Machine intelligence

In a studio behind a warehouse in western Berlin, a group of international scientists discuss our robotic future. An engineer from a large European car manufacturer issued a cautious and optimistic report on the progress of driverless cars. He explained that robot cars are learning to distinguish moving objects such as cars and pedestrians, cyclists and other relatively vulnerable. Some robot cars have been better at distinguishing different dog breeds than humans. "But of course, these are only small advances," he said.

Then, a well-dressed man wearing a light gray suit came to power. "Not long afterwards, the smartest and most important decision makers may not be humans," he said, with a smile of pity on the faces of the parents who are explaining the troubles of growing up to the teenagers. "I, we are not on the verge of another industrial revolution." But on the verge of a new type of life, it is more like a big world explosion."

Jurgen Schmid Hube is called the person who will be recognized as a father by the first autonomous robots. The algorithm developed by the 54-year-old German scientist allowed us to talk to our computer and also to allow our smartphone to translate Mandarin into English. However, he was not very keen on the future. The existence of robots will mainly serve humanity. The saying.

On the contrary, he believes that machine intelligence will soon be more than enemy intelligence, and will exceed human intelligence. The future of robots designed and manufactured by autonomous replication robotics will be closer to the sun than thin humans. They will eventually establish colonies in the asteroid belt of the Milky Way. Schmid Hubert is trying to build brains for robots.

After giving a speech, Schmid Hubert explained that in a laboratory in Lugano, in the Swiss Alps, his company, Nnaisense, has been developing a system that can operate like a baby. They have arranged small experiments for those systems. So that they understand how the world works. He referred to this as "real AI." The only problem is that they are still too slow - there are only 1 billion neural connections, and the number of neural connections in the human cerebral cortex is about 100 trillion.

"However, our current trend is that our computer can be accelerated ten times every five years, unless this trend is terminated, otherwise in 25 years we will develop a recursive neural network that rivals the human brain. The level of intelligence isn't a lot of years away, like the intelligence of a crow or capuchin."

How many years did it actually happen? "I think it's better to use the year as a unit than a decade, but I don't want to limit myself to the 4 to 7 years."

When asked why he was so confident in his timetable, he mentioned the super-light engine. We explored the revolution that jumped from the big bang to the Neolithic era, from the invention of gunpowder to the birth of the World Wide Web. Schmid Hubert pointed out that the interval between major events in the history of the universe seems to be shortening exponentially—the interval between the two milestone events is about one quarter of the previous interval. According to this law, machine intelligence beyond human intelligence seems to be occurring in 2050.

"Time will not stop by 2050, but we will usher in smarter AI than ours. At that time, the biosphere researching humanity will become meaningless. They will push history to the next stage." It will start in areas where there are abundant resources. After hundreds of years, they will establish colonies in the Galaxy."

He called this time pass "omega". The word Omega was first proposed by Teilhard de Chardin, a Jesuit pastor who was born in France in 1888. Schmid Hubert said he likes omega, "because it sounds a bit like 'oh my god'".

Controversy and criticism

Schmid Hubert's position as the godfather of machine intelligence is not entirely indisputable. As a computer scientist, sometimes his speech sounds very inconsistent with scientific principles and it is shocking. During his speech in Berlin, he could hear people complaining behind the audience. When Schmid Hubert outlined how the robot would eventually leave the earth and "go alone to enjoy" the process of exploring the universe, a neuroscientist from Brazil intervened: "This is what you have to say? The robot has to go for pleasure. The algorithm? You are destroying the scientific method in front of all these scientific people! Horrible!"

When asked about those reactions, Schmid Hube once again showed that kind of mercy. "My papers have been controversial for decades, so I'm used to these common arguments. But many neuroscientists are not really aware of what is going on in the AI ​​world."

But even in the AI ​​community, Schmid Hubert has many critics. Hearing his name, several professionals in the field of artificial intelligence have pointed out that his work is undoubtedly very influential, "it is even more and more influential," but he is "a bit angry." Many people think that his optimism about the pace of technological progress is unfounded and may even be very dangerous. Some people claimed that Schmid Hubert was far from a true prophet of robotics in the future. He actually pushed artificial intelligence to resemble the Segway balance car (when Segway was born, he was touted as a technological revolution similar to the PC invention, but eventually The desolate end of the fate.

To understand why Schmid Hubert was considered both a prophet and a joke, he must have an in-depth understanding of his resume. Born in Munich, Germany in 1963, he spent a lot of popular science books and science fiction books from a library near his home during puberty, and he became interested in robotics. His favorite books include Olaf Stapleton's Star Maker, ETA Hoffmann's The Sandman, and Lem Stanislaw. ) The novel.

He said that his most admired idol was Albert Einstein. “Sometimes, I would think if I made something smarter than myself, even something smarter than Einstein, then I might be more influential.” He obtained mathematics and computer science at the Technical University of Munich. Degree, the university granted his position as professor at the age of 30.

LSTM memory

In 1997, Schmid Hubert and his student Sepp Hochreiter co-authored a paper that proposes a way for artificial neural networks to be promoted through memory functions, that is, based on information obtained before Interpret the closed loop of the pattern of words or images. They call it the long-term memory (hereinafter referred to as "LSTM").

At the time, AI was in a long "winter" period: technology failed to reach the level of first-round touting around artificial intelligence, and financing was difficult. In the 1960s, the industry’s hope was that machines might be able to understand the complex world from the top down. If there are any new hotspots in the industry at the moment, that seems to be a relatively simple idea: Machines may be able to introduce a relative foundation but allow them to understand the algorithms of the entire complex world from the bottom up.

In 1997, Schmid Hubert's dissertation on LSTM was rejected by the Massachusetts Institute of Technology, but from now on it seems to be one of the main concepts behind the rise of deep learning. In 2015, Google announced that with LSTM, it has successfully improved the error rate of its speech recognition software by nearly 50%. LSTM is the system that drives Amazon Alexa. Apple also announced last year that it is using LSTM to improve the iPhone.

If Schmid Hubert had been successful, then the concept would be more recognized. In 2015, he complained in a harshly punished article that three Canadian computer scientists honored as superstars in the AI ​​field in Silicon Valley—Jeffrey Hinton (Google) and Jayne Lecuo (Facebook ) and Joshua Bengio (IBM) - "a large number of quotations from each other's point of view," but "has not been attributed to pioneers in the field."

During speeches and interviews in Berlin, he repeatedly stressed that the current boom in computer learning is “outdated”, and LSTM reached that point many years ago. He also feels that Silicon Valley is not so important. He feels that the region is so much dominated by "tremendous competition" that the resulting value is not as good as European research institutions.

Listening to Schmid Hubert's talk about the future of the robot, you may feel like this: He constantly talks about high-tech utopias, just to make sure that he will not end up with one of Sisto Rodriguez’s circumstances. Strategies. Rodriguez was influential, but he was not taken seriously. Instead, he entered the Hall of Honor as a Silicon Valley individual.

Will robots enslave humans?

Given that Schmid Hubert has a keen interest in science fiction, hasn't he ever feared that robots that will someday become self-conscious will enslave us and dominate us? He shook his head and said, "We will not be enslaved, at least because we are very unfit to be slaves for people who can directly create robots that are much advanced than ours." As for The Matrix, they are imprisoned. Humans were used to drive AI. He also believes that the future will not happen: “That was the stupidest plot to date. Why would you use human organisms when power stations that stock robots can generate much more power? Energy to drive the robot?"

But if that happens, will robots see it as a more efficient way to completely eliminate humanity? "Like all scientists, intelligent AI will be fascinated by the origin of life and civilization. But this enchantment will fade after a while, just like many people who do not know the origin of the world. In general, The best protection for us will be their lack of interest in us, because the greatest enemy of most species is their own people. Their attention to us will be the same as our attention to ants."

This analogy does not seem to be as satisfying as he hopes. Sometimes we inevitably step on ants? Some people even use chemicals to poison ants who live in one place. "Of course, it is only a small part of the global ant population. No one seems to want to destroy all the ants on Earth. On the contrary, most people are happy when they hear that there are more ants on Earth than humans. Most ants are in the jungle of Brazil."

He quoted a recent controversial argument from Professor Harvard Edmund O Wilson that "We may be smarter than ants, but the total human weight on Earth is almost the same as the total weight of all ants. ”

Will it cause massive unemployment?

Do not speak science fiction first. Talk about even more imminent concerns, such as the automation revolution that caused massive unemployment. In an article recently published in Nature magazine, AI researcher Kate Crawford and network law professor Ryan Calo warn that robots are used to replace humans The social chain reaction, the new wave of enthusiasm around smart design is creating a dangerous blind spot.

In this regard, Schmid Hube is not very worried. He said that when he began to give birth to two daughters in the millennium, he was very clear that the robot will come. “What advice do I give to my two daughters? I told them: Your dad thinks everything will be fine, even if it may appear ups and downs. Ready to continue to do new things. Ready to learn how to learn.”

"Humanity is always good at creating jobs that didn't exist. Most people are already doing extravagant work like your job and my job." He said, "Predict what jobs will disappear and It's not difficult, but it is not easy to predict what will happen in the future. In the 1980s, who would have thought that if someone were to become a professional video game player or a YouTube star in 30 years, they would be able to earn millions of dollars?”

“Even the respected positions in the medical community will be affected. In 2012, robots began to use deep neural networks to win cancer screenings. Does this mean that doctors will be unemployed? Of course not. It just means that The same doctor would have only had enough time to treat one patient and will be able to treat 10 patients.Many people will be able to enjoy medical research with less money.More human lives will be saved in the future. The life expectancy will also be extended."

He happily proposed that countries with more per capita robots, such as Japan, Germany, South Korea and Switzerland, will have a relatively low unemployment rate.

Will robots fall in love?

Every time you try to understand Schmid Hubert's optimistic vision for the robot's future, you will basically encounter a very simple situation. He said that when two organisms have a conflict of interest, they have two ways to resolve the conflict: either cooperation or competition.

Asked whether or not the future robot will be able to fall in love, besides being funny and fun, he gives a positive answer, saying the reason is "Love is obviously an extreme form of cooperation. The robot's emotional life will be It will be multiple and multiple, not monogamous: there will be a variety of relationships between the robots. They will be able to share their thoughts. This is something that humans cannot currently do, or only if they are opposite each other. It takes a lot of time before it is done. The robot will also have a fusion that does not exist between organisms."

If love is really just a strong form of cooperation, why does it make people feel irrational? Why do we have lust? Why do you feel broken? Schmid Hubert replied, "We already have a pain sensor, so the robot feels pain when it hits something. We will eventually understand the issue of sexual desire. After all, they all belong to the same thing."

Will the enemy state appear?

If there is a company in the future, such as Apple or Google, to form a monopoly on the world's super-smart robots, what should we do? Schmid Hubert believes that the kind of enemies "has nothing to show." In this scenario, cooperation will also be king. "The intelligent central algorithm will be very short. The algorithms that make the system self-improvement may only have 10 lines of virtual program code. The only thing we currently lack may be only 5 lines of code."

"Perhaps we will develop the 10 lines of code in my small company, but this year, even if the bank secrets of Switzerland cannot be kept, those codes will not stay in my company. They will be leaked. Perhaps India's An unknown person will think of the code and make it available to everyone."

If it sounds a bit blind and optimistic, it is because of Schmid Hubert's own experience - LSTM was initially refuted, and he has always been distrustful of "a fierce competition" in Silicon Valley - it must have taught him that competition can both create losers. Can also create winners. As far as personal feelings are concerned, his optimism will give people a sense of caution, but I will be more comfortable with the idea that the most advanced presence in the future will be facilitated by Jürgen Schmid Hube, if he is willing to clarify those Questioning words.

At the end of the interview, he apologized: "I'm sorry to make you like talking to a teenager. But I've been saying these things since the 1970s and 1980s. The only difference is that people Began to take me seriously."

Ningbo Autrends International Trade Co.,Ltd. , https://www.supermosvape.com