Connective Learning
Based on the experience of tons of failures
with a couple of successes
Dr. Virach Sornlertlamvanich
Assistant Rector for Research and Innovation, Thammasat University
SIIT, Thammasat University
Chair of Digital Cluster, RUN
virach@gmail.com, https://virach.com
KU Presidents Forum, Kaset Fair, 1 Feb 2019
“Creativity is just
connecting things.”
- Steve Jobs
KU Presidents Forum, Kaset Fair, 1 Feb 2019
Eight Challenges Innovations Confront
1. Find an idea
2. Develop a solution
The idea is one thing; a working solution is another
3. Seek sponsorship and funding
4. Reproduction
Scale up issue:- HW vs SW
5. Reach your potential customer
Wheel, steam engine, and freeze-dried foods were innovations that existed before 100
BCE, but it took centuries for innovators to position each of them in ways the average
person could use
6. Beat your competitors
The wise innovators keeps an eye on his peers’ work for purposes of collaboration,
inspiration, or tactical recognizance
7. Timing
Why automobiles are rated in horsepower and electric lights in candles. As great as your
idea is, will the culture be ready when it’s finished?
8. Keep the lights on
The myths of Innovation, Scott Berkun, O’Reilly, 2007
Learning
KU Presidents Forum, Kaset Fair, 1 Feb 2019
Paradigm Shift in
Research and Education
KU Presidents Forum, Kaset Fair, 1 Feb 2019
Lesson Learned from
the Past Research Strategy
Many successes in Research,
but very few can be Commercialized.
WHY?
Business
Academ
ic
The key is Problem Setting
- What topics that can serve both Research
interest and Industrial needs?
- How the Research topics can be realized
for Industrial needs?
- How the Industrial needs can be
interpreted to Research interest?
Boost up
Research and
Innovation
KU Presidents Forum, Kaset Fair, 1 Feb 2019
Issues
• Mismatch between research interest and industrial
needs
• Unmatched skill sets
• Research vs. Development
• Prototyping vs. Production
• Proving vs. Operating
• Lack of proper business model for the research i.e.
licensing, pricing, etc.
• Lack of experience of supporting team, PI, PM, PC, etc.
Resulting in “On the shelf” vs. “Off the shelf” research
KU Presidents Forum, Kaset Fair, 1 Feb 2019
R&D 1.0
R&D Industry
R&D 2.0
R&D Industry
R&D 3.0
R&D Industry
R&D 4.0
R&D Industry
Research Industry Result
Output oriented Technology sourcing Useless
- Market oriented
- High price
- Requirement
unmatched
- High cost
- Discontinuous dev.
and service
- Re-development
- Not ready to use
- Unusable research
Targeted budget Cost reduction Development only
- Development
oriented
- Lots of fund
- Avoid internal
research
- Fund acquirement
- Superficial dev
- Low research work
Co-research Research privilege R&I value added
- TM
- STEM
- Incubation center
- Co-proposal
- Save cost
- Tax exemption
- STI product
- High value product
Research Business No collaboration
- Researcher driven - Market driven
- OEM
- On the shelf
- No R&D product
(expecting)
Research Industry Result
Output oriented Technology sourcing Useless
- Market oriented
- High price
- Requirement
unmatched
- High cost
- Discontinuous dev.
and service
- Re-development
- Not ready to use
- Unusable research
Targeted budget Cost reduction Development only
- Development
oriented
- Lots of fund
- Avoid internal
research
- Fund acquirement
- Superficial dev
- Low research work
Co-research Research privilege Share cost
- TM
- STEM
- Incubation center
- Co-proposal
- Save cost
- Tax exemption
- Superficial proposal
- Impractical
Research Business No collaboration
- Researcher driven - Market driven
- OEM
- On the shelf
- No R&D product
(probably)
R&D 1.0
R&D Industry
R&D 2.0
R&D Industry
R&D 3.0
R&D Industry
R&D 4.0
R&D Industry
R&D 4.0
R&D Industry
R&D Trap
2010
1980
1990
2000
KU Presidents Forum, Kaset Fair, 1 Feb 2019
We need a design of
connective learning
KU Presidents Forum, Kaset Fair, 1 Feb 2019
Require and Deliverable
University Industryrequire
Research
Innovation
Market
Researcher
RA
deliverable
Publication
Patent
Prototype
Research
Innovation
Graduate
Revenue
deliverable
Product
Business
Patent
Revenue
KU Presidents Forum, Kaset Fair, 1 Feb 2019
Require and Deliverable
University Industryrequire
Research
Innovation
Market
Researcher
RA
deliverable
Publication
Patent
Prototype
Research
Innovation
Graduate
Revenue
deliverable
Product
Business
Patent
Revenue
Non-STI
product
Unused
research
Research and Industry Mismatch Problems
Problem is not that the research is not good, or researchers do not perform,
but it is the problem of research authorities who do not know how to make use of it.
KU Presidents Forum, Kaset Fair, 1 Feb 2019
Require and Provide
University Industryrequire
Budget
Equipment
Topic
provide
Research
Innovation
Consultancy
Network
Research
RA
provide
Outsource
Equipment
Topic
Staff
require
Research
Innovation
Market
Researcher
RA
deliverable
Publication
Patent
Prototype
Research
Innovation
Graduate
Revenue
deliverable
Product
Business
Patent
Revenue
CRDC-SIIT (2016-2018)
Center for Research and Development Collaboration
KU Presidents Forum, Kaset Fair, 1 Feb 2019
CRDC-SIIT as a Role of Catalyst (2016-2018)
Center for Research and Development Collaboration @SIIT
KU Presidents Forum, Kaset Fair, 1 Feb 2019
•Faculty member
•Researcher
•Policy execution
•Funding agency
•Co-research space
•Course work
•Seminar, training
•Business partner
•Investor
Business
Partnership
Innovation
hub
Network of
Professional
Support of
Policy and
Sourcing
BINS
GD.FINDI
(Production Line Simulator)
• J Industry LEXER
• T Industry DHAS TPA
• Univ. SIIT, TNI
Seminar ConsultancyR&D Education Business
KU Presidents Forum, Kaset Fair, 1 Feb 2019
MAT2016
Compact EV
• J Industry FOMM
• T Industry Bangchak
• Univ. SIIT
Evaluation R&D Business
Thai PBS
2016/1/31
KU Presidents Forum, Kaset Fair, 1 Feb 2019
Mimamori System
(Elderly Care
Support System)
• J Industry AIVS
• T Industry Banphaeo
Hospital, Sawangkanives
• Univ. SIIT
Evaluation R&D Business
KU Presidents Forum, Kaset Fair, 1 Feb 2019
DTAC AI Lab
• Established in 2017
• Sponsor 8-10 research projects
• To connect data science
students to leading professors
with real-world business cases.
• Initially, focus on time-
consuming processes best left
to machines
KU Presidents Forum, Kaset Fair, 1 Feb 2019
SIIT-LINE Developer Challenge
• It is the first hackathon that
Line provides specifically to SIIT
students
• It is a programming and
business model competitions
• Duration: 4-months (Aug-Dec
2018)
• During the program, Line
provides Line APIs and
mentoring & tutorial sessions
for SIIT students to develop the
applications
• Based on 2018, 9 teams (27
students) joined the
competition
KU Presidents Forum, Kaset Fair, 1 Feb 2019
AI Boom
Opportunity and Risk
KU Presidents Forum, Kaset Fair, 1 Feb 2019
What is AI?
--classic and modern aspects--
• From a behavioral point of view, is an artificial
agent that shows certain characteristics of
intelligence like:
• Perception
• Knowledge acquisition
• Knowledge representation
• Reasoning
• Planning
ó Regression
ó Deep learning
ó Modeling
ó Prediction
ó Recognition
KU Presidents Forum, Kaset Fair, 1 Feb 2019
Differences within AI
Artificial Intelligence
• General AI
• Vertical AI (Expert Systems)
• Natural Language Processing
• Computer Vision
• Machine Learning
• ...
KU Presidents Forum, Kaset Fair, 1 Feb 2019
Thinking, Fast and Slow
by Nobel laureate Daniel Kahneman on AI (2011)
--The two systems--
http://upfrontanalytics.com/market-research-system-1-vs-system-2-decision-making/
KU Presidents Forum, Kaset Fair, 1 Feb 2019
Multilayered neural
networks to vast amounts
of data
Enable machines to
improve at tasks with
experience
Mimic human intelligence
using logic, if-then rules,
decision trees, machine
learning and deep learning
Deep Learning (Neural learning from data with high quality, but imperfect results)
Watson (Associative learning from data with high quality, but imperfect results)
Semantic Web (Knowledge graph links formation from extraction, clustering and learning)
Modern AI is making some huge strides
KU Presidents Forum, Kaset Fair, 1 Feb 2019
AI advancement that brings about the 3rd AI Boom
• Thinking Machines
• DeepBlue Chess Machine (1997)
• IBM Watson Quiz Show (2011)
• DeepMind AlphaGo (2016)
• Self-Driving Cars
• RHINO Museum Tour Guide (1997)
• DARPA Grand Challenge (2005)
• Google Self-driving Car (2011)
• Smart Assistants
• Apple Siri Personal Assistant (2011)
• Amazon Echo & Alexa (2014)
• Google Home & Assistant (2016)
Byoung-Tak Zhang, “Human-Level AI and Video Turing Test”
Google’s AlphaGo AI narrowly beats the
world’s top human Go player 2017
SIliconangle
Geospatialworld
Pocket-lint
KU Presidents Forum, Kaset Fair, 1 Feb 2019
Action
Execution
Control
Input
Acquisition
Recognition
Comparable AI Frameworks
Data Processing Storage Analysis
AI x RobotAI x IoT
Visualization
Sensing Recognition Modeling Planning Action
Accumulating Knowledging Understanding Solving
Data Visualization
KU Presidents Forum, Kaset Fair, 1 Feb 2019
How can AI be our
educational partner?
Elderly care robots Fujitsu Pepper
JRA
CyberDyne
HRP
Cinnamon
Panasonic
Aibo
Asimo
KU Presidents Forum, Kaset Fair, 1 Feb 2019
Learning AI
— Online Self-Learning MOOC Style —
https://futuremonger.com/artificial-intelligence-learning-phd-or-mooc-selflearning-458059725421
KU Presidents Forum, Kaset Fair, 1 Feb 2019
Lecture Note Generation
(VDO Summarization)
KU Presidents Forum, Kaset Fair, 1 Feb 2019
Summary
Multilingual support
VDO Playback
Educational Partner
for self-learning
Scene search
Hyper link
Word Cloud
KU Presidents Forum, Kaset Fair, 1 Feb 2019
Speech-to-Text Translate Summarize
Keyword
Extraction
Lecture Note
Word Cloud
Subtitle
Synchronize
Subtitles
System Architecture
KU Presidents Forum, Kaset Fair, 1 Feb 2019
Challenges
• Data -> Annotated Data -> Cooked Data
• Data -> Metadata -> Hyper Data
Open, Share and Connect
In concern with:
• Data Accessibility
• Privacy
• Security
• Intellectual Property
KU Presidents Forum, Kaset Fair, 1 Feb 2019
Challenges
• Current AI is nothing more than a machine that has a
capability to learn.
• AI should not only be able to learn and reason, it should also be able
to interact and react.
• AI platforms should do more than answer simple questions.
They should be able to learn at scale, reason with purpose,
and naturally interact with humans. They should gain
knowledge over time as they continue to learn from their
interactions, creating new opportunities for business and
positively impacting society.
• AI will result in net job gain. Reskill for new job role to work
with AI
KU Presidents Forum, Kaset Fair, 1 Feb 2019
Summary 1/2
• Lesson learned from the past research strategy
• Mismatch between research interest and industrial
needs
• Unmatched skill sets
• Lack of proper business model for the research i.e.
licensing, pricing, etc.
• Lack of experience of supporting team, PI, PM, PC, etc.
• Critical factors for success in university-industry
research projects
• Technology
• Corporate usefulness
• Corporate capacity
• Corporate confidence
KU Presidents Forum, Kaset Fair, 1 Feb 2019
Summary 2/2
• Challenge in CRDC’s University-Industry Link
• From Require and Deliverable model
to Require and Provide model
• CRDC-SIIT as a role of catalyst
• Business partnership
• Innovation hub
• Network of professional
• Support of policy and sourcing
• Win-win university-industry collaboration
• Research faculty
• Market potential
• STI value added product
• Research facility
• Capacity building
• Research publication
KU Presidents Forum, Kaset Fair, 1 Feb 2019

Connective Learning

  • 1.
    Connective Learning Based onthe experience of tons of failures with a couple of successes Dr. Virach Sornlertlamvanich Assistant Rector for Research and Innovation, Thammasat University SIIT, Thammasat University Chair of Digital Cluster, RUN virach@gmail.com, https://virach.com KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 2.
    “Creativity is just connectingthings.” - Steve Jobs KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 3.
    Eight Challenges InnovationsConfront 1. Find an idea 2. Develop a solution The idea is one thing; a working solution is another 3. Seek sponsorship and funding 4. Reproduction Scale up issue:- HW vs SW 5. Reach your potential customer Wheel, steam engine, and freeze-dried foods were innovations that existed before 100 BCE, but it took centuries for innovators to position each of them in ways the average person could use 6. Beat your competitors The wise innovators keeps an eye on his peers’ work for purposes of collaboration, inspiration, or tactical recognizance 7. Timing Why automobiles are rated in horsepower and electric lights in candles. As great as your idea is, will the culture be ready when it’s finished? 8. Keep the lights on The myths of Innovation, Scott Berkun, O’Reilly, 2007 Learning KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 4.
    Paradigm Shift in Researchand Education KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 5.
    Lesson Learned from thePast Research Strategy Many successes in Research, but very few can be Commercialized. WHY? Business Academ ic The key is Problem Setting - What topics that can serve both Research interest and Industrial needs? - How the Research topics can be realized for Industrial needs? - How the Industrial needs can be interpreted to Research interest? Boost up Research and Innovation KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 6.
    Issues • Mismatch betweenresearch interest and industrial needs • Unmatched skill sets • Research vs. Development • Prototyping vs. Production • Proving vs. Operating • Lack of proper business model for the research i.e. licensing, pricing, etc. • Lack of experience of supporting team, PI, PM, PC, etc. Resulting in “On the shelf” vs. “Off the shelf” research KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 7.
    R&D 1.0 R&D Industry R&D2.0 R&D Industry R&D 3.0 R&D Industry R&D 4.0 R&D Industry Research Industry Result Output oriented Technology sourcing Useless - Market oriented - High price - Requirement unmatched - High cost - Discontinuous dev. and service - Re-development - Not ready to use - Unusable research Targeted budget Cost reduction Development only - Development oriented - Lots of fund - Avoid internal research - Fund acquirement - Superficial dev - Low research work Co-research Research privilege R&I value added - TM - STEM - Incubation center - Co-proposal - Save cost - Tax exemption - STI product - High value product Research Business No collaboration - Researcher driven - Market driven - OEM - On the shelf - No R&D product (expecting)
  • 8.
    Research Industry Result Outputoriented Technology sourcing Useless - Market oriented - High price - Requirement unmatched - High cost - Discontinuous dev. and service - Re-development - Not ready to use - Unusable research Targeted budget Cost reduction Development only - Development oriented - Lots of fund - Avoid internal research - Fund acquirement - Superficial dev - Low research work Co-research Research privilege Share cost - TM - STEM - Incubation center - Co-proposal - Save cost - Tax exemption - Superficial proposal - Impractical Research Business No collaboration - Researcher driven - Market driven - OEM - On the shelf - No R&D product (probably) R&D 1.0 R&D Industry R&D 2.0 R&D Industry R&D 3.0 R&D Industry R&D 4.0 R&D Industry R&D 4.0 R&D Industry
  • 9.
    R&D Trap 2010 1980 1990 2000 KU PresidentsForum, Kaset Fair, 1 Feb 2019
  • 10.
    We need adesign of connective learning KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 11.
    Require and Deliverable UniversityIndustryrequire Research Innovation Market Researcher RA deliverable Publication Patent Prototype Research Innovation Graduate Revenue deliverable Product Business Patent Revenue KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 12.
    Require and Deliverable UniversityIndustryrequire Research Innovation Market Researcher RA deliverable Publication Patent Prototype Research Innovation Graduate Revenue deliverable Product Business Patent Revenue Non-STI product Unused research Research and Industry Mismatch Problems Problem is not that the research is not good, or researchers do not perform, but it is the problem of research authorities who do not know how to make use of it. KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 13.
    Require and Provide UniversityIndustryrequire Budget Equipment Topic provide Research Innovation Consultancy Network Research RA provide Outsource Equipment Topic Staff require Research Innovation Market Researcher RA deliverable Publication Patent Prototype Research Innovation Graduate Revenue deliverable Product Business Patent Revenue CRDC-SIIT (2016-2018) Center for Research and Development Collaboration KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 14.
    CRDC-SIIT as aRole of Catalyst (2016-2018) Center for Research and Development Collaboration @SIIT KU Presidents Forum, Kaset Fair, 1 Feb 2019 •Faculty member •Researcher •Policy execution •Funding agency •Co-research space •Course work •Seminar, training •Business partner •Investor Business Partnership Innovation hub Network of Professional Support of Policy and Sourcing BINS
  • 15.
    GD.FINDI (Production Line Simulator) •J Industry LEXER • T Industry DHAS TPA • Univ. SIIT, TNI Seminar ConsultancyR&D Education Business KU Presidents Forum, Kaset Fair, 1 Feb 2019 MAT2016
  • 16.
    Compact EV • JIndustry FOMM • T Industry Bangchak • Univ. SIIT Evaluation R&D Business Thai PBS 2016/1/31 KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 17.
    Mimamori System (Elderly Care SupportSystem) • J Industry AIVS • T Industry Banphaeo Hospital, Sawangkanives • Univ. SIIT Evaluation R&D Business KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 18.
    DTAC AI Lab •Established in 2017 • Sponsor 8-10 research projects • To connect data science students to leading professors with real-world business cases. • Initially, focus on time- consuming processes best left to machines KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 19.
    SIIT-LINE Developer Challenge •It is the first hackathon that Line provides specifically to SIIT students • It is a programming and business model competitions • Duration: 4-months (Aug-Dec 2018) • During the program, Line provides Line APIs and mentoring & tutorial sessions for SIIT students to develop the applications • Based on 2018, 9 teams (27 students) joined the competition KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 20.
    AI Boom Opportunity andRisk KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 21.
    What is AI? --classicand modern aspects-- • From a behavioral point of view, is an artificial agent that shows certain characteristics of intelligence like: • Perception • Knowledge acquisition • Knowledge representation • Reasoning • Planning ó Regression ó Deep learning ó Modeling ó Prediction ó Recognition KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 22.
    Differences within AI ArtificialIntelligence • General AI • Vertical AI (Expert Systems) • Natural Language Processing • Computer Vision • Machine Learning • ... KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 23.
    Thinking, Fast andSlow by Nobel laureate Daniel Kahneman on AI (2011) --The two systems-- http://upfrontanalytics.com/market-research-system-1-vs-system-2-decision-making/ KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 24.
    Multilayered neural networks tovast amounts of data Enable machines to improve at tasks with experience Mimic human intelligence using logic, if-then rules, decision trees, machine learning and deep learning Deep Learning (Neural learning from data with high quality, but imperfect results) Watson (Associative learning from data with high quality, but imperfect results) Semantic Web (Knowledge graph links formation from extraction, clustering and learning) Modern AI is making some huge strides KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 25.
    AI advancement thatbrings about the 3rd AI Boom • Thinking Machines • DeepBlue Chess Machine (1997) • IBM Watson Quiz Show (2011) • DeepMind AlphaGo (2016) • Self-Driving Cars • RHINO Museum Tour Guide (1997) • DARPA Grand Challenge (2005) • Google Self-driving Car (2011) • Smart Assistants • Apple Siri Personal Assistant (2011) • Amazon Echo & Alexa (2014) • Google Home & Assistant (2016) Byoung-Tak Zhang, “Human-Level AI and Video Turing Test” Google’s AlphaGo AI narrowly beats the world’s top human Go player 2017 SIliconangle Geospatialworld Pocket-lint KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 26.
    Action Execution Control Input Acquisition Recognition Comparable AI Frameworks DataProcessing Storage Analysis AI x RobotAI x IoT Visualization Sensing Recognition Modeling Planning Action Accumulating Knowledging Understanding Solving Data Visualization KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 27.
    How can AIbe our educational partner? Elderly care robots Fujitsu Pepper JRA CyberDyne HRP Cinnamon Panasonic Aibo Asimo KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 28.
    Learning AI — Online Self-LearningMOOC Style — https://futuremonger.com/artificial-intelligence-learning-phd-or-mooc-selflearning-458059725421 KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 29.
    Lecture Note Generation (VDOSummarization) KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 30.
    Summary Multilingual support VDO Playback EducationalPartner for self-learning Scene search Hyper link Word Cloud KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 31.
    Speech-to-Text Translate Summarize Keyword Extraction LectureNote Word Cloud Subtitle Synchronize Subtitles System Architecture KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 32.
    Challenges • Data ->Annotated Data -> Cooked Data • Data -> Metadata -> Hyper Data Open, Share and Connect In concern with: • Data Accessibility • Privacy • Security • Intellectual Property KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 33.
    Challenges • Current AIis nothing more than a machine that has a capability to learn. • AI should not only be able to learn and reason, it should also be able to interact and react. • AI platforms should do more than answer simple questions. They should be able to learn at scale, reason with purpose, and naturally interact with humans. They should gain knowledge over time as they continue to learn from their interactions, creating new opportunities for business and positively impacting society. • AI will result in net job gain. Reskill for new job role to work with AI KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 34.
    Summary 1/2 • Lessonlearned from the past research strategy • Mismatch between research interest and industrial needs • Unmatched skill sets • Lack of proper business model for the research i.e. licensing, pricing, etc. • Lack of experience of supporting team, PI, PM, PC, etc. • Critical factors for success in university-industry research projects • Technology • Corporate usefulness • Corporate capacity • Corporate confidence KU Presidents Forum, Kaset Fair, 1 Feb 2019
  • 35.
    Summary 2/2 • Challengein CRDC’s University-Industry Link • From Require and Deliverable model to Require and Provide model • CRDC-SIIT as a role of catalyst • Business partnership • Innovation hub • Network of professional • Support of policy and sourcing • Win-win university-industry collaboration • Research faculty • Market potential • STI value added product • Research facility • Capacity building • Research publication KU Presidents Forum, Kaset Fair, 1 Feb 2019