Shulou(Shulou.com)06/02 Report--

There is a psychological phenomenon in psychology that if any word is repeated enough times, the word will lose all its meaning. If such a concept had been used to describe artificial intelligence, the word artificial intelligence should have fallen apart a long time ago. However, the reality is that artificial intelligence appears more frequently, is ubiquitous in the field of science and technology, and its significance has never diminished.

To be exact, artificial intelligence has been misused and is so hyped and touted in the commercial market that it is difficult for many people to identify with it. at the same time, there are some exaggerated studies in the media reports. label any vague AI story as "reckless".

Many people often sum up whether artificial intelligence is intelligent or not. For non-professionals, this can be a thorny issue. Compared to ordinary people, they are more willing to confuse it with the most familiar scientific vision: a conscious computer that is many times smarter than humans. Experts call this concrete example of artificial intelligence artificial general intelligence, and if we really want to create such a thing, there is still a long way to go. Before that, no one would benefit from exaggerating the wisdom or capabilities of artificial intelligence systems.

In spite of this, it still does not prevent this technology from contributing more to our society than ever before. Today's medical care, the management of enterprises, the development of smart cities. It helps people create music and books, it can read your resume carefully so that your resume can be seen by as many companies as possible, and solve the problem of whether the sales-oriented enterprise phone robot is easy to use. Or, the increase in your credit card points and the adjustment of photos on your phone. Whether you can accept it or not, it is changing the way you live.

So, instead of talking about artificial intelligence, talk about "machine learning". This is a sub-field of artificial intelligence, covering almost all the research methods that have the greatest impact on the world. As such a word, it does not have the mystery of artificial intelligence, but it is more helpful to explain the role of the technology.

There are many explanations for machine learning, but in essence, the computer itself "learns". But it will bring a bigger problem.

Let's start with a question. Suppose you want to create a program that can identify dogs. You can try to do it in traditional ways, such as using clear rules such as "dogs have protruding teeth" and "dogs have fluffy tails". But what will the program do if there is a picture of a wolf? So you need to program every necessary rule, but this is very time-consuming, and you have to define various difficult concepts in the process, such as "teeth" and "tail". Then the best option is to let the program learn on its own. If you think that recognizing pictures is too common, and if you want to recognize voices, especially human voices, it is best to be able to talk in real time, then this involves the industry of intelligent voice, although the direction is different, but it is also the corresponding principle. For example, smart phone robots, which are different from traditional smart speakers, are collected by big data and integrated on this basis to build a complete voice system.

In fact, the biggest advantage of this approach is the most obvious: you never need to write a program. Of course, you need to do a lot of patching, improve the way the system processes data, and find smarter ways to get this information, but you don't tell the program what it should be looking for. This means that it can find patterns that humans may miss or never thought of. Because all a program needs is data, you can train it for any job. After all, modern society is full of data. With the hammer of machine learning in hand, even if the digital world is full of nails, you can hammer it into the right place at any time.

Instructing computer self-study is a wonderful shortcut. Like all shortcuts, it needs to be opportunistic. If you like, you can think that there is intelligence in artificial intelligence. But this is not true wisdom in the human sense, nor does it follow human principles. You may wonder: how smart can a book be? What expertise can be contained in a pan?

So where are we now in terms of artificial intelligence? After years of headlines announcing the next major breakthrough, some experts think we have reached some kind of calm. But this is not a real obstacle to progress. In terms of research, there are a large number of ways to explore in our existing knowledge, but in terms of products, we only see the tip of the iceberg of the algorithm.

Lee describes the current era as an "era of execution", in which technology begins to "spread from the laboratory to the world". Benedict Evans, another venture capital strategist, likens machine learning to a relational database. Relational database is a kind of enterprise software that made a lot of money in the 1990s and revolutionized the industry. But machine learning is so common that your eyes may simply see the word. What these two are trying to say is that we are now in the stage of a rapid return to normal artificial intelligence. "in the end, almost everything will be machine learning somewhere inside it, but no one will care," Evans said. "

He's right, but we're not there yet.

For now, artificial intelligence-machine learning-is still a new thing, often unexplained or understudied. But in the future, artificial intelligence will be so common that it will be ignored by you.

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.