Smart furniture
cleaning robot
intelligent speaker
Personal assistant
Translator
Google translate
Microsoft translate
YouTube video translation
image recognition
Face Recognition
AI + camera
Chess master
Alphago 2017 as the world's first defeat of Kejie Cheng
Alphago Go 0-100 lost Alphagozero
artificial intelligence has been involved with the game StarCraft DOTA2
Autopilot
Google's Waymo
autonomous vehicles van
Andrew Ng Drive.ai
Baidu unmanned vehicles
ready to mass production
of medical and health
Deep Blue II Waston IBM
DeepMind AlphaGo have created a department DeepMind health
medical robot gradually popularity of
financial and business
intelligence investment adviser
Capricorn investment
consonance investment
intelligence service
AI referral service
financial regulation
What is artificial intelligence
OCR image processing
without learning is not AI
artificial intelligence are constantly learning among
AI classic problem
Turing test
Chinese House
first wave of
1949, the first available EDVAC computer programming computer program stored in binary
computer von Neumann architecture of the
first winter 1970
AI bottleneck can only solve the problem in the easiest part
performance limited limited memory and processing speed can not solve complex problems exponentially
lack of "common sense" many AI applications listed as machine vision and natural language need a lot of basic understanding of the world's
second wave
expert system stores a large number of domain-specific knowledge and experience
DEC's commercial systems and expert system developed at the University CMU XCON
Lisp language functional programming language
single-user workstations
Symbolics
fifth-generation project have not seen the British investment AI
Hopfield new network
of new training algorithm for
second winter
Apple, IBM began to promote the first-generation desktop machine
property issues market needs are replaced by government stopped funding the desktop
expert system performance issues very practical limitations of
the third wave of
the late 1990s has
dark blue Deepblue 1997 victory over world champion Garry Kasparov (chess rules-based search , pruning)
based on a large opening book, library statistics final valuation results
Chess vs Go
rules are much simpler but huge chessboard
large data
machine translation (statistical models) for translation based on statistical methods
AlphaGo Ke Jie and Li Shishi has won a
powerful utility can into life using the
social needs of the community increasing demand for
artificial intelligence development of
today's artificial intelligence in the end how smart?
AI will develop in the end to what extent?
(human tool that can be utilized) weak artificial intelligence focus on solving a specific problem areas artificial Intelligence in particular collar Beyond human technology tools available for human use
strong universal AI Artificial Intelligence (does not appear quite comprehensive and humans)
With representation and integration of knowledge, learning ability reasoning ability to communicate the overall ability to achieve the set goals through natural language
controversy: whether they have human consciousness
super artificial intelligence (beyond the range of human perception) in scientific creativity, wisdom and every aspect of social skills and other than the most intelligent human brain lot of intelligent
speed up or bottlenecks
Moore's Law every year and a half computing power will double the prices will fall by 50%