Ke Jie vs Alpha Dog, lost
Recently, Ke Jie, who represents the highest level of human Go, and the "Alpha Dog" who swept the Korean player Li Shishi last 4:1 in the first game of the game in Tongxiang, Zhejiang. After a 4.5-hour battle between the two sides, the Go AI AlphaGo, developed by the Deepmind team, defeated the Chinese chess player Ke Jie, who ranked first in the world, and won the top spot.
Compared with AlphaGo, which defeated Korean player Li Shishi last spring, Ke Jie’s version of Artificial Intelligence 2.0 is obviously more powerful today. After the game, he said that he had done his best, but the opponent was crazy and long. "I hope to do my best to fight every game. I think his game at this time is exactly two people last year. The first time His chess is still close to humans. Now his understanding of chess is simply Go God."
The results of the doctors' group and AI competition in 7 hospitals in Beijing...
In April, Beijing Friendship Hospital brought out one of the big AI children: a thyroid nodule ultrasound image diagnosis AI, with the Beijing Union Hospital, Peking University Third Hospital, Beijing Cancer Hospital, People's Liberation Army General Hospital and other seven top three hospitals The video doctor of the war fought a battle.
There are a total of 100 questions in the game, and the choice of "benign" or "malignant" is judged according to the image. Most of the questions contain malignant and suspicious points, which are difficult. In the second half, each half is limited to 20 minutes. A total of 84 doctors participated in the competition. In addition to the formation of 7 groups by the hospital, there were also individual participants. At the same time, there are nearly 10,000 doctors watching the game online, and if you want, you can also play online.
In the first part of the answer process, the top 25 doctors in the accuracy rate took an average of 1,000 seconds, while the AI ​​took 191.669 seconds, the correct rate was 68%, while the primary physician's correct rate was 60.8%, and the intermediate physician's was 62.4%. Senior physician 66%. In the second part of the question, the AI ​​accuracy rate was 76%, and only 5 doctors scored more than it.
In the end, in the results of the team competition, AI won with a comprehensive correct rate of 73.0%, taking 514 seconds, leading the second space center hospital 0.3%.
In his personal performance, AI ranked sixth. The champion belongs to Dr. Zhang Mingbo from the Department of Ultrasound, General Hospital of PLA. She has 8 years of experience in reading. The comprehensive correct rate is 76%, and the average time is 909 seconds.
For the results obtained by AI, the doctors of the Ultrasound Department of Beijing Friendship Hospital are still quite satisfied: the final AI score is 83.3% sensitivity and 57.5% specificity.
The game shows that human diagnosis rate may be fluctuating, not as stable as AI.
According to reports, the Artificial Intelligence Summit Forum was held in Wuzhen, China as scheduled. At the summit forum, a group of Google executives made a collective appearance, telling the research and development process of Alpha Go and the future development of the AI ​​field for the on-site participants and industry professionals.
In the roundtable discussion session, Chen Gang, deputy dean of the School of Computer Science, Zhejiang University, Jeff dean, senior researcher of Google brain+, Liu Zhiqing, director of the Computer Go Game Research Institute of Beijing University of Posts and Telecommunications, and Li Jia, director of R&D of Google AI and cloud, were invited. Attendance, and discussed the current development of AI and future trends, and published their own views.
Talking about the status quo: people and AI complement each other and have their own advantages
When talking about whether AI's wisdom can transcend human problems, Chen Gang believes that there should be a process of complementarity and synergy between people and AI. Jeff said that AI is superior to humans in some fields today, but compared with people, AI is still in the field of learning human habits, and has not yet reached the ability of human beings to have autonomy.
Li Jia believes that AI has a long life compared with people. If you let humans do more complex processes, there will be multiple problems, but let robots with AI ability can cope with them and face behavior. For example, writing poetry and interactive dialogue to people, there is still a gap between AI robots. In the long run, AI and people will play their respective strengths and make the whole society better.
As a domestic expert who also studied the field of Go, Liu Zhiqing talked about his views from the perspective of Go. He believes that human intelligence and AI machines are fundamentally different and cannot be compared. For example, Alpha Go, which we saw yesterday, is predicting the winning rate when playing chess, while professional players use point-by-point methods. Therefore, from the perspective of Go, AI and human wisdom are complementary, and they can learn from each other and help each other because they are all winning the game.
Talking about the future: education, medical care or industry into the next blue ocean
For the future development of the AI ​​industry, it will be in the field of human society in addition to the field of Go, and the guests who participated in the roundtable discussion also gave different views. Chen Gang said that the education field needs the most intervention of AI, because in the traditional teaching behavior, a teacher will formulate a unified teaching plan for multiple students, and then understand the degree of knowledge each student receives, which greatly affects efficiency. After the introduction of AI machine-assisted instruction, it is possible to accurately determine the depth of knowledge that each student receives and to achieve personalized teaching.
Google brain+ senior researcher Jeff dean
Jeff believes that the future of the medical field will become the next main battlefield for AI. He believes that now we go to the hospital to see a doctor, doctors need to rely on their own experience to judge the disease, but the doctor's own experience is limited, if relying on the assistance of AI robots, bringing together the treatment plan given by all doctors in the world under the same condition, Then the doctor can switch from a confirmed diagnosis to a small-scale screening of the treatment plan, which improves the efficiency of the consultation. However, Jeff stressed that in this process, the role of AI robots is to replace the doctor's diagnosis, not the process of consultation.
Li Jia said that overall, by combining the knowledge and experience accumulated by human beings for many years, the combination of human and AI machines can collide with more sparks. In addition to the medical field, in the field of agriculture, AI robots are also promising.
Liu Zhiqing said that AI has made breakthroughs in the field of Go, providing a good solution for human beings in the field of natural resources and transportation. Just like Go, through the accurate positioning of each move, to effectively use natural resources and improve logistics distribution efficiency, ease urban traffic congestion, improve logistics and distribution efficiency, the future AI is promising.
AI has not yet "cultivated into fine"
Interestingly, when the round-table dialogue is about to end, whether the AI ​​robot will suddenly have a sense of autonomy in the future, Jeff said that every move that Alpha Go takes is done under the precise algorithm given by humans. However, it is not certain that AI robots can generate autonomy. "But what we do is to ensure safety in the first place, so this is a software problem. For example, in the process of Go, we often see Alpha Go making some orders. Go players are incredibly good, which is exactly what we want Alpha Go to achieve, and Alpha Go itself creates the expected results," he said.
Regarding whether AI robots will have "motivation" behavior, Chen Gang said with a joke. "First of all, consciousness itself is a philosophical problem, and it has not yet reached the point of scientific behavior, so I think it is difficult to say that robots have their own consciousness."
Li Jia bluntly said that it is too early to say that robots have their own "motivation". She believes that human beings now use AI robots as an auxiliary tool and are still in the early stages of AI exploration. Liu Zhiqing also agrees with Li Jia's point of view, but adds that there is no scientific basis at present to believe that robots can generate consciousness on their own.
step 2: Immediately close the container again.
This protects the remaining test strips from humidity and guarantees strip integrity up to the expiry date indicated.
step 3: Dip the test strip for about 1 second into the fresh urine specimen.
Wipe off any excess urine on the rim of the vessel and blot the edge of the test strip on tissue paper.
step 4: Read the result by comparing the test fields with the color scale on the test pack after 30 to 60 seconds.
Urine Test Strips 2K,Urinalysis Strips,Best Urine Test Strips,Urine Analysis Strips
Changchun LYZ Technology Co., Ltd , https://www.lyzstrips.com