Ai titech-virach-20191026
- 1. When AI becomes a Data-driven
Machine!!
Virach Sornlertlamvanich
Professor, AAII, Musashino University
Chair of Digital Cluster, RUN
SIIT, Thammasat University
National Distinguished Researcher Award 2003, NRCT
virach@gmail.com
The 12th Thai Kuramae Kai General Meeting, Disruptive Technology Seminar, TNI, October 26, 2019
- 2. “The development of full artificial intelligence could
spell the end of the human race ….It would take off
on its own, and re-design itself at an ever increasing
rate. Humans, who are limited by slow biological
evolution, couldn't compete, and would be
superseded.”— Stephen Hawking
“Artificial intelligence will reach human levels by
around 2029. Follow that out further to, say, 2045,
we will have multiplied the intelligence, the human
biological machine intelligence of our civilization a
billion-fold.” — Ray Kurzweil, author
“The pace of progress in artificial intelligence is
incredibly fast. ... The risk of something seriously
dangerous happening is in the five-year timeframe.
10 years at most.” — Elon Musk, CEO of Tesla
28 Best Quotes About Artificial Intelligence, 2018
Pessimists
- 3. “Using up data and AI is only a means to survive.”
— Kenichiro Yoshida, Sony President
“Some people call this artificial intelligence, but the
reality is this technology will enhance us. So instead
of artificial intelligence, I think we'll augment our
intelligence.” — Ginni Rometty, CEO of IBM
28 Best Quotes About Artificial Intelligence, 2018
“Artificial intelligence would be the ultimate version
of Google. The ultimate search engine that would
understand everything on the web. ...” — Larry
Page, CEO of Alphabet
Optimists
- 4. 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
“It’s a big thing to integrate [causality] into AI,” Bengio says. “Current
approaches to machine learning assume that the trained AI system will be
applied on the same kind of data as the training data. In real life it is often
not the case.” [Deep learning is blind to cause and effect]
- 5. Differences within AI
Artificial Intelligence
• General AI
• Vertical AI (Expert Systems)
• Natural Language Processing
• Computer Vision
• Machine Learning
• ...
- 6. Thinking, Fast and Slow
by Nobel laureate Daniel Kahneman (2011)
--The two systems--
http://upfrontanalytics.com/market-research-system-1-vs-system-2-decision-making/
- 7. AI advancement that brings about the 3rd AI Boom
• Thinking Machines
• DeepBlue Chess Machine (1997)
• IBM Watson Quiz Show (2011)
• DeepMind AlphaGo (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
• Self-Driving Cars
• RHINO Museum Tour Guide (1997)
• DARPA Grand Challenge (2005)
• Google Self-driving Car (2011)
Pocket-lint• Smart Assistants
• Apple Siri Personal Assistant (2011)
• Amazon Echo & Alexa (2014)
• Google Home & Assistant (2016)
- 9. 1,000,000 Programmers needs in 2037
http://www.thansettakij.com/2017/03/08/133452
Thailand 4.0
Data Science
Big Data
Text
Analytics
Fintech
Machine
Learning
Deep
LearningKnowledge
Science
Language
Engineering
IoT
Data Mining
Data
Surveillance
Artificial
Intelligence
Robotics
Autonomous
Vehicle
NLP
Machine
Translation
• 50,000 programmers (2017)
• 100,000 programmers (needed in 2017)
• 6,000 programmers/year (graduated)
• 2,000 programmers (qualified)
Recommendation (Accenture’s Future Workforce, 2018):
- Needs of new skilling to work with intelligent machines.
- Map new skills to new roles.
- 11. Three Big AI Research Institutes
• AIRC by AIST, Japan
• 2015
• Focusing on translational research
• Researchers: 77 (2015) -> 400++ (2017)
• 20-50 members in each research team
• DFKI, Germany
• Since 1988
• 900 researchers (510 employees), 80 spin-offs
• World largest AI research center
• USC, US
• Data Science Platform
- 12. AI
Ontology /
Knowledge
Simulation
/ Multi-
agent
Machine
Learning
AIRC/AIST, Japan
Sensing Recognition Modeling Planning Action
Inference
HPC for AI
AI x RobotAI x IoT
Data acquisition
Recognition
Action, Planning
and Execution
Prof. Dr. Jun-ichi Tsujii
Target areas
1. Mobility
2. Productivity
3. Healthcare, Welfare
4. Safety, Security
Development in 3-layers for collaboration
L3: Shared Tasks and Benchmark Data
Geo, Life, Robot, Science
L2: AI Framework and Advanced Modules
Data acquisition, Recognition, Planning, NLP
L1: Large-scale fundamental Research
Machine Learning, Probabilistic, Brain-inspired
AI, Data, Knowledge
- 13. DFKI, Germany
German Research Center for Artificial Intelligence
(Deutsches Forschungszentrum für Künstliche Intelligenz)
• PPP/JV on AI
• Develop Open Platform for
• Setting up network for industry and research
• Academia
• Industry
• Collaboration framework
• Digital reality to scale AI
• CERN of AI
Prof. Dr. Philipp Slusallek
- 14. Data Science Institute (DSI), USC, US
• Data science platforms
DSI
Data Platform
Societal impact
Research publication
Technology transfer
Real world
problem
Prof. Dr. Cyrus Shahabi
- 17. 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.
• Deep learning is good at finding patterns in reams of data, but can't
explain how they're connected. (Turing Award winner Yoshua Bengio)
• AI will result in net job gain. Reskill for new job role to work with
AI