Display: block; max-width: 100%; margin: 0px auto; word-wrap: break-word !important;” alt=”Lee Sedol: I will never play chess with "Alpha Dog" again” />
In the man-machine battle of Go held in March this year, AlphaGo finally defeated the famous Korean player Lee Sedol 9 dan 4:1, which attracted widespread attention from all over the world, and also brought concepts such as neuroscience and deep learning into the public consciousness. vision.
This year’s Davos also held a symposium dedicated to the battle of the century to explore the technological and social implications of computer programs defeating the world champion of Go.
The meeting invited Li Shishi, two artificial intelligence experts and scholars, Dileep George, co-founder and chief technology officer of Vicarious, and Wendell Wallach, a scholar at the Interdisciplinary Center for Bioethics at Yale University, to participate in the discussion. The host of the meeting was Li Mei, Dean of the School of Entrepreneurship and Management, ShanghaiTech University.
This sub-forum attracted the attention of many participants. There was a long queue 30 minutes before the start. In the end, apart from the participants, only 5 media were allowed to enter the venue to listen. Jiemian was fortunate to be one of these 5 media. .
At the beginning of the meeting, the host Li Mei threw questions to the guests on the definition of “artificial intelligence”, and the guests had different opinions.
Dileep George believes that the ability to understand and understand the world through data and master problem-solving skills through different learning models is called “artificial intelligence”. And Wendell Wallach believes that being able to learn from data is a major feature of artificial intelligence, but because machines are currently unable to perceive things in complex environments, there is still a certain distance to achieve true artificial intelligence.
The one who really has an intuitive understanding of “artificial intelligence” is undoubtedly Lee Sedol, who directly confronted AlphaGo. When asked about his feelings against AlphaGo at that time, Lee Sedol said that his biggest feeling was “surprise”.
“Go has many uncertain factors, the limitations of the machine are many, and loopholes will also appear in the process of playing chess, so watching the game between AlphaGo and Fan Hui before made me think that the computer cannot reach the highest level of Go players. The result Everyone knows: I lost.” Lee Sedol recalled.
Since people will show more or less emotions when they play games, facing AlphaGo who has no emotions, it also puts more psychological pressure on Li Shishi, which makes it difficult.
“People will have psychological sway, even if they know the exact answer, they may still choose another path and consider other options at the next moment. But Alpha Dog will not have any sway, this is what I face The biggest difficulty.” Li Shishi said that the game between man and machine will be in a very unfavorable situation. For example, if there are three games, the players will usually use different openings for each game, while AlphaGo will not change.
“It’s very uncomfortable to face an opponent who has no emotions. It makes me feel that I don’t want to play with it anymore.” Li Shishi reluctantly admitted that if there is another chance to play against AlphaGo, the situation will definitely be very difficult.
However, in the face of hundreds of CPUs, GPUs and AlphaGo supported by many scientists, Li Shishi did not feel unfair. He believed that his biggest mistake was that he did not have enough understanding of the field of deep learning and despised his opponents.
“After watching the game between AlphaGo and Fan Hui, I always thought I would win, but I was surprised that AlphaGo’s chess skills have improved so much in just six months.” Before, Li Shishi would do a lot of drills in his mind, but because the opponents on the court were not human, it made him feel very unfamiliar.
Looking back on those few games, Li Shishi believed that he still had a chance to defeat AlphaGo. He felt that due to his misjudgment of AlphaGo’s strength, he did not play his due strength. “But who knows how it has grown in the past few months?” Li Shishi still expressed his doubts.
AlphaGo has had out-of-the-box performances in competitions. According to the definition of human beings, this may be the “creativeness” shown by the machine.
However, the three guests all believed that the definition of “creativity” for people and machines is different, because the cognitive scope of the two sides is not the same. “AlphaGo may not even know he’s playing chess,” Lee Sedol said.
After AlphaGo’s impressive achievements, what impact will artificial intelligence have on our lives? Wendell Wallach believes that this shows that machine learning capabilities are increasingly important, such as the ability of machines to recognize images, It has brought people applications in face recognition, unmanned driving, weather prediction, etc. Deep learning will also have a wide range of applications in the medical field.
Wendell Wallach also pointed out that artificial intelligence will face some dilemmas in terms of social ethics. For example, in the face of an impending traffic accident, should driverless artificial intelligence save passengers or people on the road? People surveyed generally agreed that they should make the choice of “saving more lives”, but were asked whether they would buy For such a driverless car, the answer for most people is not to buy it. This shows that people still have distrust of artificial intelligence.
Li Shishi believes that the future development of artificial intelligence in the field of Go is much simpler: “Artificial intelligence is developing, but people are also developing.” He believes that although it will become more and more difficult for humans to defeat artificial intelligence in the field of Go, because artificial intelligence is Predictable and able to complete his tasks with great accuracy, he is willing to let artificial intelligence guide his daughter to learn Go.
In the end, the host Li Mei concluded that people often put artificial intelligence on the opposite side of human beings. For example, people think that people’s jobs will be replaced by machines, but in fact, they cultivate the next generation, let human beings grow up with technology, and match human capabilities with Only when the capabilities of machines are integrated can super intelligence be generated and more problems can be solved.
The Links: 3HAC042475-001 3HAC033498-001