For beginners to learn AI sometimes need a little background, I gather from the Internet simply summarized under to others.
First, the history of the development of artificial intelligence
Development of artificial intelligence has not been easy, the general was "three from two down" trend, now regarded as a new era towards artificial intelligence development.
Start (1) Dream (1900--1956). In 1900, at the Congress of Mathematicians Hilbert solemnly announced the 23 unsolved problem to the world of mathematicians. The second question and the tenth question in this 23 problem and is closely related to artificial intelligence, and ultimately promote the invention of the computer. Turing conceived out according to the tenth issue of the Turing machine, which is the theoretical models of the computer, the successful portrayal of the meaning of the mechanized operation process, and ultimately pave the way for the invention of the computer. In 1954, von Neumann completed early design of EDVAC computer, and proposed a "von Neumann architecture." Overall, Turing, Godel, von Neumann, Wiener, Claude Shannon and other great pioneers laid the foundation of artificial intelligence and computer technology.
(2) The Golden Age (1956--1974). In 1965, McCarthy, Minsky and other scientists organized by the "Dartmouth Conference", first proposed the "AI (Artificial Intelligence, referred to as AI)" concept marks the birth of artificial intelligence disciplines. Thereafter, the artificial intelligence research into the golden age of 20 years, have made a number of remarkable research achievements, such as machine theorem proving, checkers procedures, set off the first wave of development of artificial intelligence. In this golden age, John McCarthy developed the LISP voice, decades later become the most important field of artificial intelligence programming language; Min Marvin Chomsky have a more in-depth study of the neural network, also we found a simple neural network is insufficient; and more layer neural network, back-propagation algorithm began to appear; expert system also started.
(3) first AI winter - Reflections on Development (1974--1980). Early breakthrough artificial intelligence development greatly enhance people's expectations of artificial intelligence, overvalued the pace of development of science and technology. However, the fall spate of failures and expectations of the development of artificial intelligence into a valley. In 1973, Lighthill report on artificial intelligence, artificial intelligence, opened the prelude to winter. Since then, the scientific community for a round of in-depth artificial intelligence of torture, the AI is subjected to harsh criticism and questions about its real value. Subsequently, governments and institutions have stopped or reduced funding, artificial intelligence into its first winter in the 1970s. Limited computing power, the lack of data makes a lot of sense development into a bottleneck, especially over-reliance on computing power and experience of the amount of data neural network technology, a long time did not achieve substantial progress, especially after the "Perception," a book published on the nervous network technology produced a devastating blow, further research into almost nobody in the subsequent decade. Expert System at the end of this era emerged and opened the next era.
( 4) Application Development (1980--1987). Expert system simulation knowledge and experience of human experts to solve problems in specific areas of artificial intelligence to achieve practical application of theoretical study on the steering using expertise from the general reasoning strategy of a major breakthrough. Expert system to succeed in the medical, chemical, geological and other fields, and promote a new wave of artificial intelligence into the application development. 1980 Carnegie Mellon University (CMU) developed XCON officially put into use, which became a milestone in the new era, the beginning of an expert system power play in specific areas, but to the entire artificial intelligence technology has entered a boom phase. After 10 years of silence, the neural network has made new progress in research, we found that the neural network algorithm has the ability to learn, which makes it all the way to the development of neural networks, in the back of the 1990s commercial, is used for character image recognition and speech recognition .
(5) Second AI winter - depressed Development (1987--1993). With the application of artificial intelligence continues to expand the scale, there is the field of application of expert systems narrow, lack of common sense knowledge, knowledge acquisition difficulty, single reasoning, the lack of a distributed function, it is difficult to be compatible with existing databases and other problems gradually exposed. Field of artificial intelligence was mainly used John McCarthy's LISP programming language, PC progressive development of LISP machines are booming beat, dedicated LISP machine hardware sales serious market collapse, artificial intelligence once again into the winter. Confused rout and theoretical studies of the hardware market, coupled with Governments and institutions have stopped invest in research into artificial intelligence, it led to several years of low, but it also made some important achievements. In 1988, American scientist Judea Peel the probabilistic reasoning methods to introduce artificial intelligence in which the later development of artificial intelligence has played a significant impact. IBM's Watson Research Center, the probabilistic method introduced to the language processing artificial intelligence; in 1992, Kai-fu Lee to use statistical methods, design and development of the world's first speaker-independent continuous speech recognition program; in 1989, AT & T Bell labs Jahn of Music and Queensland team uses convolution neural network technology, artificial intelligence, digital image of the postal code identify the handwriting.
(6) the steady development (1993--2011). Since the development of network technology especially Internet technology, accelerate the innovative research in artificial intelligence, artificial intelligence technology to promote further into practical use. In 1995, Richard Wallace developed a new chat bot Alice, it can make use of the Internet continues to increase its own set of data, optimize content. In 1997, IMB's Deep blue computer Deep Blue defeated the human world chess champion Garry Kasparov. In 1997, the German Science Huokelaite and施米德赫伯proposed long-term memory (LSTM) This is a recurrent neural network is still used today, one kind of handwriting recognition and speech recognition, have a profound influence on subsequent research in artificial intelligence . In 2004, the American neuroscientist Jeff Hawkins book "The Future of Artificial Intelligence," a book put forward a new theory to predict the brain and memory, according to this theory point out how to build truly intelligent machine, the book of the later nerve scientific-depth study had a profound impact. In 2006, Jiefulixin Dayton published "Learning Multiple Layers of Representation" later laid a new architecture of neural networks, artificial intelligence technology is still the core of the depth of learning.
(7) New Era (2012-- date). With mobile Internet technology, the outbreak of cloud computing technology, the amount of data accumulated in the history of the imagination, which provides enough power for the material and the subsequent development of artificial intelligence, neural networks, represented by the depth of the rapid development of artificial intelligence techniques significantly across the "technology gap" between science and application, to usher in a new wave of explosive growth. . Artificial intelligence, big data, cloud computing, networking technology, together constitute the second decade of the 21st century technology theme. In 2012, the depth of convolution neural network algorithm design of the visual identity challenge by the University of Toronto in ImageNet held, is the industry that is beginning to learn the depth of the revolution. In 2014, Ian Gudefeiluo against the proposed GANs generation network algorithm, which is a kind of artificial intelligence algorithms for unsupervised learning, the algorithm consists generation networks and network assessment, this approach quickly Artificial Intelligence many areas of technology adoption. 2016 and 2017, Google launched a sensation in the world of two man-machine war Go, its artificial intelligence program AlphaGo consecutive victory over former world champion Go Shishi Korea, as well as the current world chess champion China's Jie Ke, caused a huge stir. Speech recognition, image recognition, and other unmanned technology deepening.