The History of Artificial Intelligence - From Turing Test to Big Data

Read the full text[url]
http://click.aliyun.com/m/23311/[/url]
I have always been curious about how artificial intelligence was proposed, what kind of story is behind it, and in these 60 years of artificial intelligence development What happened during the years? Why is now the breaking point of artificial intelligence, and where will artificial intelligence go in the future? With this question in mind, I read Dr. Wu Jun's book "The Age of Intelligence", which opened up my understanding of artificial intelligence. The main content of this article also comes from this book.

Our generation's attention to artificial intelligence comes from the 2016 AlphaGo battle against the world-famous Go player Li Shimin. Before the game, the attention of all parties was very high, and various domestic media rushed to report and predict the result of this game. People are curious about artificial intelligence. How smart is the intelligence now and how does the computer play Go with people? In the end, AlphaGo defeated Li Shiming 4:1. Everyone is feeling that the era of artificial intelligence is coming. Just one year later, on May 27, 2017, AlphaGo's version 2.0 defeated Ke Jie, the world's No. 1 Go player, 3:0. Since then, there has been no human opponent in front of AlphaGo.

The reason why computers can beat humans is because machines acquire intelligence in a different way than humans. It does not rely on logical reasoning, but on big data and algorithms. Google trained AlphaGo using data from hundreds of thousands of games played between master Go players, which is how it acquired its so-called "intelligence." In terms of computing, Google used hundreds of thousands of servers to train the AlphaGo chess model and let different AlphaGos play tens of millions of games against each other. The second key technology is the heuristic search algorithm - Monte Carlo tree search algorithm (English: Monte Carlo tree search; referred to as: MCTS), which can limit the search space to a very limited range and ensure that the computer can quickly find good the next method. It can be seen that the seemingly intelligent problem of playing Go is, in essence, a problem of big data and algorithms.

When it comes to artificial intelligence, we have to mention a legend in the computer industry: Dr. Alan Turing. In 1950, Turing published a paper "Computing Machines and Intelligence" in the journal Mind. In the paper, Turing did not talk about how computers can acquire intelligence, nor did he propose an intelligent method to solve complex problems. Knowledge proposed a method to verify whether a machine has intelligence or not.



Let a machine and a person sit behind the scenes, and let a referee communicate with the person and the machine behind the scenes at the same time. If the referee cannot judge whether the person he is communicating with is a human or a machine, it means that the machine has the same intelligence as a human. . It is the famous Turing test. Later, computer scientists added that a computer can be considered to have the kind of intelligence Turing called it if it achieves one of the following things:

1. Speech recognition
2. Machine translation
3. Automatic textualization Abstract or write
4. Chess champions beat
humans