Business and AI
- 1. AI and Business AI could be our last invention Shinichiro Yano
- 2. The goal of this lecture • General trend on intelligence • What is intelligence? • How far ultimately AI can reach? • AI will take over human? • What have been studied so far? • What is happening now, • Where are we now? • What would happen in 10 years? • How business could change? Don’t confuse those two! HAL I, Robot
- 3. Disclaimer • Predicting future is not possible Our universe, Our life Never repeat Artificial life You can not repeat process Our universe 1.38*10^10 year old, would probably exist 10^10^26?? If it were 1 years, Christ, Budda, Mohamed were born last few seconds, our knowledge just expand in our universe quite recently
- 4. What is intelligence? • Definitions • Individuals differ from one another in their ability to understand complex ideas to adapt effectively to the environment to learn from experience, to engage in various forms of reasoning to overcome obstacle by taking thought. (American Psychological Association task force on Intelligence) • They are not quantitative difference, they are qualitative difference Intelligence person could see stuff normal people don’t • F = T 𝛻 S, Drive the system to maximize future freedom of action “Intelligence should be viewed as a physical process that tries to maximize future freedom action and avoids the constraints its own future.” PRL 110 168702 (2013) Intelligence explored because human invents tools and start using them. AI could be last tool human will invent
- 5. Intelligence • They are not quantitative difference, they are qualitative difference • They see stuff we don’t János Mattis-Teutsch Kouhei Uchimura Richard Feynman
- 6. Proposition • Our brain activities can be realized by computer -> There is no reason we can’t simulate thinking, recognition, memory, feelings Neural network
- 7. Life is advancing • Biological • Evolve hardware • Evolve software • Cultural • Technological • Evolve hardware • Design software • Design hardware • Design software
- 8. 3 components of our cognitive activities • What do we think • How you feel • What do you believe - Pathos Emotions/Values Ethos Credibility/Trust Logos Logic/Reason/Proof Aristotle (350 B.C.)
- 9. Immanuel Kant (1724-1804?) • Critiques of pure reason • What do we think • Metaphycis • Critiques of judgement • How you feel • Ascetics, theory of beauty • Critiques of practical reason • What do you believe • Ethics Judgement Practical Pure reason Immanuel Kant (1724-1804)
- 10. Artificial intellingence can’t believe • They could potentially think and feel but not believe • Our cognition is active (Construcuralism) • Our understanding is based on experience and perception, we can extend out cognitive activity with metaphysical activities. • Our understanding = unification of our judgement on things = not necessary true at all. • While we have a lot to answer our cognitive activities. Metaphysical activities here Hardest for computer. Sensing /Observation Understanding /Meaning Integration Judgement/Execution
- 11. What will bottlenecks be Design goals Understanding Structure Proposition Execution It does not have will Sensors are not developed like humans Needs data It can not have questions Cant design framework No sense of common sense Can not order human
- 12. It is very far. AI on economy Human level AI Superintelligence Who is in charge? Humans Cyborgs AI How will humans treated? Well Replaced Badly
- 13. Anyways, What is matter now??
- 14. What is happening now? • 1st Reasoning (1956- 1960-) • 10 years after 1st computer • Decision making tree,Game theory • Chess, Go • 2nd Expert system (1980-) • 5th generation computer (Japan) • Dark age after this era • 3rd Machine learning(2010 -) • Web + Big data • Computation power • Breakthrough with Deep learning 3G AI ≠ neural network
- 15. What is happening now? • Recognition • Information ……………..Web search, movie, music, picture, voice. • Voice, picture, movie …….. Feeling, cancer detections, movie editing • Detecting anomaly …….. Fraud, suspicious person or trade, disaster • Prediction • Number ………….. Sales, economic parameters, election, cancer, credit • Demands ………… order, users preferences, • Matching ……. Advertising, recommendations, • Execution • Presentation ………… Summary, Writing, Translations, Composition, Drawing • Design …………………. Chart, Logo, Site Design, Medicine, Architecture, Cooking • Optimization ………... Game, Transportation, Place for new stores, medical • Automation ………….. Q&A, Mobile, car, SEO, Cooking, operations
- 16. It is very far. Imaging Multimodal Planning/action Abstraction Language Understanding Understand language Acquire knowledge through language Understand feelings 2030?? 2020??
- 17. We enter all connected age Telecommunication • 1G phone • 2G mail • 3G i-mode • 4G smartphone • 5G ??? <- wearable? IoT Business • B-B • B-B -> B-C • B-C • B-C, C-C • B-B-C, C-B-C How the system could change with post-smarthone? How the money flow changes by that? Computer • IBM • Microsoft • Google • Apple, Facebook, Amazon • ??
- 18. 4 Key technologies backed by 5G • IoT • Smartphone, wearable devices Smart home, driverless car • Blockchain • ANNGOUKA, Consensus algorism P2P, Distributed Ledger Technology • Cloud • Amazon AWS, IBM, Microsoft • AI (3rd generations) • Analyses and uses data Collect data Store data Manage data Analyse data
- 19. What are jobs?? • Job is about solving problem • So you need to find problem which people or technology can’t solve easily • Job is about entreating people • AI can’t do anything meaningless
- 20. Job wont be gone but parts of job are automated • Medical • Legal system • Office • IT • Publishing • Sales • Mobile • Real-estate • Security Ex.)
- 21. Entertainment wont be gone but part of they are connected, depend, and attended • Sports • Show • Fashion • Music • Events Ex.)
- 22. We enter imaginary dimensions • We should play more in imaginary dimensions Re (実数) Im (虚数) Centralized Decentralized Distributed Problem Design Mass production Product Problem AI A product Solution problem -> theory -> experiment -> solution Problem -> data -> AI -> solution -> theory Problem solving Production Open and close Grab
- 23. We enter imaginary dimensions • We should play more in imaginary dimensions • But remember the 4 questions, to evaluate business What value do they provide?To whom? How do they deliver? How do they make money?