At present, the discussion about artificial intelligence is in full swing, and in May, Google Go artificial intelligence "AlphaGo" will stage a "man-machine war" with China chess player Ke Jie, one of the strongest Go players in Zhejiang Wuzhen. Everyone is curious: will artificial intelligence be more powerful than human intelligence?
Before answering this question, we should understand the development status of artificial intelligence, explore the future development direction and prospects, and look at the "golden age" of artificial intelligence from a calm perspective.
— — Editor
A few days ago, at the CITE2017 Artificial Intelligence Industry Development Summit Forum held in Shenzhen, Liu Lihua, Vice Minister of the Ministry of Industry and Information Technology, said that China has been at the forefront of the world in artificial intelligence (AI) technology research, and researchers and entrepreneurs in this field "stand at the forefront of the global tide". In the future, China has very valuable opportunities and prospects in this field.
Recently, the topic of artificial intelligence has not diminished, and the IT field even calls it. At what stage does artificial intelligence develop at present? Is it really powerful enough to challenge human intelligence? How should we treat the so-called artificial intelligence boom?
The golden age of artificial intelligence development has come.
"The recent achievements of artificial intelligence are really unexpected for more than a decade." Liu Tieyan, chief researcher of Microsoft Research Asia, believes that today is a "golden age" of artificial intelligence.
From the technical breakthrough of artificial intelligence, in terms of voice and image recognition, the level of computers has been similar to or even surpassed that of human beings in specific fields and categories. For example, in December last year, Microsoft released a new live translation function of Microsoft Translator (Microsoft Translation APP), which supports more than 50 languages and can realize real-time translation of up to 100 people in multiple languages. People from different countries can communicate with each other by using this APP with their mobile phones, say a word or enter words, and what they hear or see is each other’s mother tongue.
In addition, artificial intelligence has also made a breakthrough in the cognitive field that is traditionally considered difficult to be successful by machines. For example, when training artificial intelligence to play a game of knocking bricks, artificial intelligence found a very effective way to score after 120 minutes. After training for 240 minutes, it reached the level of "hardcore" players, and even discovered some tricks that human players did not find at ordinary times.
Shen Xiangyang, global executive vice president of Microsoft, and Li Yanhong, chairman and CEO of Baidu, have experienced the low tide of artificial intelligence development. When studying artificial intelligence in the 1990s, this course was considered to have no business prospects. "There are three main reasons why artificial intelligence is on fire today: a large amount of data on the Internet, powerful computing power, and breakthroughs in deep learning." Shen Xiangyang said that deep learning is one of the machine learning methods, and it is a way for computers to learn from the surrounding world or a specific example and become more intelligent. Li Yanhong said that with the development of the Internet, the data is getting richer and richer, the computing resources are getting cheaper and cheaper, and the breakthrough of deep learning in the past five or six years has laid the foundation for the development of artificial intelligence.
It will take a long time to really challenge people’s cognitive ability.
Although the development is in full swing, will artificial intelligence be smarter than human beings in the future? Nearly half a century’s research shows that machines have incomparable advantages in search, calculation, storage and optimization, but they can’t compete with humans in perception, reasoning, induction and learning.
"It will take a long time or even impossible for artificial intelligence to really challenge people’s cognitive ability. But when it can approach human beings, it will gradually subvert various industries. " For example, Li Yanhong said that the current airport security check needs to compare ID cards. When face recognition develops faster and faster, security check may be more intelligent and convenient. In the future, people don’t need to learn how to use tools, just tell them what to do.
"People are stupid, but the machine is more stupid. In the history of artificial intelligence, from the 1960s to the 1980s, people conducted rule-oriented artificial intelligence for the first time and finally failed. The second time, people handed the answer to the machine, but finally found that the machine still could not understand the world; Now is the third time, human beings put ‘ Routine ’ Teach the machine. " Wang Xiaochuan, CEO of sogou, believes that "artificial intelligence is still in a very early stage and it is impossible to replace human beings. Anxiety is completely unnecessary."
So, will artificial intelligence be by going up one flight of stairs on the current basis? In addition to the current commercial application-oriented research schools, scientists in the field of basic research are also trying to promote the progress of artificial intelligence through hybrid intelligence. This kind of hybrid intelligence connects biological intelligence and machine intelligence to create a more powerful intelligent form, which not only has the ability of environmental perception, memory, reasoning and learning of human agents, but also has the ability of information integration, search and calculation of machine agents. Ma Huateng, Chairman and CEO of Tencent, believes that the current artificial intelligence is confined to a relatively narrow field and belongs to a very limited skill simulation, but it is possible to realize general artificial intelligence in the next step.
Facing the upsurge, we need to look on coldly and respond positively.
At the just-concluded 2017 China (Shenzhen) IT Leaders Summit, some experts pointed out that no country in the world has artificial intelligence as extensive as China, from intelligent robots to intelligent medical care, intelligent security, smart homes and smart cities, to voice recognition, gesture control and autonomous driving, but the artificial intelligence platform built by top domestic Internet companies is still far behind Amazon and Google.
"The surface of domestic artificial intelligence looks very hot. In fact, if this magic tablecloth is unfolded, you will find that it is riddled with holes and all kinds of very basic problems have not been solved." In Liu Tieyan’s view, "it is more meaningful to calmly examine the nature of technology and escort the sustainable development of artificial intelligence."
Liu Tieyan said that from the historical experience, every rise of artificial intelligence is due to some advanced technological invention, and every time a bottleneck is encountered, it is because people’s expectations for artificial intelligence are too high, which exceeds the level that technology can achieve. This will lead to divestment, and researchers will not have enough funds to engage in research.
Although there are lessons of failure, many people are still optimistic about the future of artificial intelligence. "Artificial intelligence is a very big industry that will last for a long time and will develop rapidly in the next 20 to 50 years." Li Yanhong judged, "Under the tide of this era, not a company can do everything, but it is necessary to promote the development of the whole artificial intelligence technology together." Shen Xiangyang believes that in the future 5-mdash; In 10 years, the artificial intelligence in perception will rapidly progress to the level that can rival human intelligence, and computer voice and vision will even surpass people. In the longer term, every commercial application may be subverted by artificial intelligence.
How should China deal with the changes caused by artificial intelligence? Zhu Min, president of Tsinghua University National Finance Research Institute, believes that in this round of changes in artificial intelligence, China and the world are on the same starting line for the first time, and they also have corresponding capital, companies, technology and talents for the first time. Governments, enterprises and individuals should cherish the opportunity, participate in this great change and strive to lead this round of changes. Specifically, the government should fully understand the challenges and changes of artificial intelligence, cultivate the foundation and environment for the development of artificial intelligence, support basic research, and discuss, establish and develop the ethical standards of artificial intelligence; Enterprises can’t just stay in the product market, but also move down to be a platform and a foundation; Individuals must pay attention to the cultivation of artificial intelligence skills and improve their ability to understand data.