Shulou(Shulou.com)06/03 Report--

It has been three years since beating Lee se-dol 4-1 in the AlphaGo go tournament in March 2016. In recent years, under the restraint of Google and other large companies, the word "artificial intelligence" is not uncommon. With the end of the man-machine war, the "artificial intelligence" tuyere has also been detonated to the extreme. You can see "smart" projects everywhere: smart voice, smart speakers, smart phone robots, smartwatches (bracelets), various smart homes, and so on.

Today, artificial intelligence is the product of a breakthrough in "deep learning" (no wonder the arithmetic department is so hard to squeeze in). Deep learning algorithm is essentially a kind of neural network, which was proposed as early as the 1940s, and reached a research upsurge in the 1980s.

If we want to make a deep correction of the "deep learning" algorithm, it is only adding more hidden layers of the neural network, which is essentially a neural network. The mathematical basis of "neural network" algorithm is probability theory. To put it simply, the speech recognition algorithm finds the closest text in the corpus; the autopilot algorithm calculates the safest driving route; when AlphaGo plays chess, it always calculates the most likely to win, and so on.

If we only look at artificial intelligence from the algorithmic level, then what the current artificial intelligence uses is only the upgrade at the algorithmic level, and there is no difference between the core essence and the ability to think independently compared with the 1980s. So why is artificial intelligence so different from what it is now? For example, phone robots are gradually replacing customers in the human-computer interaction experience. The key point is still big data.

As early as 1980, the famous futurist Toffler enthusiastically put forward the concept of "big data" in his work "the third Wave" and called it "the Chinese color movement of the third wave", but it was not until June 2011 that McKinsey, a world-famous management consulting company, published a big data report, which was also big data's first practical application.

Nowadays, the word "big data" is often mentioned with "artificial intelligence". It also has a very tacky name-"A lot of data". For example, user data (web pages, items browsed or purchased by users, and even the content of some chat software), traffic data, weather data, stock market data, etc., with the surge in the number of Internet users and the improvement of server storage computing power, storage and analysis of "a lot of data" has become a reality.

Because the word "a lot of data" is needed in commercial publicity, it cannot be compared with the word "big data" at all.

"big data" is a literal translation from English "big data". If this concept was first invented in China, it might be called "Hongdata". It would feel different.

According to the current recognition of the industry, artificial intelligence has three levels:

I. Computational intelligence

Fast computing and memory storage capacity.

Second, perceptual intelligence

Visual, auditory, tactile and other perception. Both humans and animals have the ability to interact with nature through a variety of intelligent perception. Self-driving cars realize such perceptual intelligence through perceptual devices such as lidar and artificial intelligence algorithms.

III. Cognitive intelligence

Popularly speaking, it is "able to understand and think". Only when human beings have language, can they have concepts and reasoning, so concepts, consciousness and ideas are all the manifestations of human cognitive intelligence.

In fact, the above two intelligences are realized at the software level, and the man-machine interface is easier for people to use, but there is still a gap between them and the real intelligence. But the power of business is terrible and can change the meaning of a word! Nowadays, the word "intelligence" is almost equivalent to "easy to use". If a product is changed to be easy to use, it makes perfect sense and is easier to understand, but it is not as high-end as the word "intelligence", and it is more valuable for commercial promotion without the word "intelligence".

Similar to so-and-so furniture, under the banner of artificial intelligence, it is actually a frame answer under a fixed program, and even the highest principle of technology is only similar to that of induction lights.

We know that "artificial intelligence" in recent years is a business word, but there is really not much we can do. After all, it is also a game played by bosses, and as onlookers, we can't make people forget the word "artificial intelligence". However, we can strengthen our resistance to this word by ourselves. If words such as "artificial intelligence" home appear in the selection and purchase of goods, it depends on whether it is really what we need or whether it is just marked with this label.

It can be said that there is no essential breakthrough from the Turing machine-von Neumann architecture to now. If it is progress, it should be the advances in chips and 5G that have improved the computing power of AI, just like the smart phone robot mentioned above. Whether a phone robot is easy to use depends on the data and system computing power.

Although the current upsurge of "artificial intelligence" does not live up to its name, the help given to some enterprises by information technology is real, especially those with external calls, which can bring a lot of convenience. This is a good time, at least we are still looking forward to a breakthrough in basic physics. The completeness of quantum, string theory or M theory and the creation of a new branch of mathematics. The purpose of this paper is to provide theoretical support for the realization of intelligent algorithms different from probability and a breakthrough in brain biology. Thoroughly understand the structure of the human brain, why it produces consciousness, and establish an accurate mathematical model for it. Breaking the emergence of the Turing machine-von Neumann architecture computer is the time to realize the real artificial intelligence.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.