Electronic enthusiasts eight o'clock in the morning: On the day of the mathematics college entrance examination on June 7, two special "candidates" participated in the college entrance examination. One of the “candidates†in Beijing took less than ten minutes to complete the 2017 Beijing Volume Arts Mathematics Question and completed the entire test paper including objective questions and subjective questions with a score of 134 points. It is said that this is the result of the candidate's six-fold slowdown in the speed of the question. In Chengdu, two thousand kilometers away from Beijing, another “candidate†took 22 minutes to complete the same exam with a score of 105.
The two candidates are intelligent AI robots, which are the intelligent education robot Aidam developed by Xuebajun and the artificial intelligence system AI-Maths developed by Chengdu Zhuxing Xingxue Technology Co., Ltd.
Some people believe that artificial intelligence will become the most important solution for the transformation of the education industry in the next 5-10 years. Promoting intelligent education and building a learner-centered teaching model can improve the shortcomings of traditional education and make up for the shortcomings of traditional education. board.
Why does intelligent AI choose to take math?
The intelligent education robot Aidam developed by Xuebajun and the artificial intelligence system AI-Maths developed by Chengdu Zhuxing Xingxue Technology Co., Ltd. all selected the 2017 college entrance examination math test.
Why are intelligent robots keen on competing mathematics?
Zhang Kailei, the founder of Xuebajun, said, "One thousand people have a thousand Hamlet in their hearts, but there is only one Pythagorean theorem in the world." Mathematics, physics and other science subjects are strong logical links. The test sites are basically stable and easy to judge right or wrong. At the same time, the mathematics papers include simple multiple-choice questions, as well as complex questions that require a problem-solving process, which is very suitable for testing AI.
Zhang Kailei revealed that in the experimental stage, Aidam's best score was 139 points, and the usual stable score was 120-130 points.
Chen Ruifeng, the chief scientist of Xuebajun, said at the test site that Aidam's problem solving is divided into three major steps: First, the test paper title becomes a formal language that the machine can understand. Second, after a complete description of a series of environments in a formal language, the computer initiates a search in the knowledge network, confirms which knowledge points the title triggered, and then performs a series of inferences to find the best solution path. Third, transform the process and answer of the problem from formal language to natural language and complete the problem-solving process.
Unlike Aidam, AI-MATHS is another way of thinking. Lin Hui, CEO of Chengdu Zhuxing Yunxue Technology Co., Ltd. and director of the Big Data Center of Suzhou Research Institute of Tsinghua University, told the media that AI-MATHS solves the problem through a comprehensive logical reasoning platform instead of learning to store the question bank, so it completely breaks the question bank. AI-MATHS scored 93 points in February this year, under the constraints of a small sample of 12 servers and 12,000 questions.
Ren Linhui revealed that before February of this year, AI-MATHS had only 100 sets of test questions. From February onwards, in order to train the speed of the AI-MATHS, the team increased the number of questions. Currently, AI-MATHS has 500 sets of questions. The accumulation.
In contrast, it is possible that AI-MATHS is closer to the autonomous learning method of artificial intelligence complex logic reasoning.
Intelligent AI's "short board" in reading comprehension
Although AI robots can make rapid progress in mathematics, there is still a way to become a student.
At the same time that the intelligent education robot Aidam did the mathematics for the college entrance examination, the three sets of mathematics college entrance examination champions were on the same stage as the PK. The scores of the three groups of college entrance examinations were 146 points, 140 points and 119 points respectively, and Aidam scored 134 points. The three groups of college entrance examinations scored an average of 135 points, and Aidam lost by one point.
AI-MATHS scored 105 points for the 2017 liberal arts mathematics in the 2017 college entrance examination, which took 22 minutes. In 2017, the national college entrance examination for the liberal arts in the college entrance examination won 100 points and took 10 minutes. For the results of AI-MATHS, Lin Hui, CEO of the Visionary Clouds, said in an interview with the media that the important reason for its loss is that it is "incomprehensible."
In January of this year, researchers at the National Institute of Informatics (NII) announced that they had abandoned the plan to let the artificial intelligence system "Torobo-kun" participate in the entrance examination of the University of Tokyo. Professor Noriko Arai of NII explained: “The artificial intelligence system cannot understand the necessary information, and the ability to read and understand the meaning of sentences has limitations. We have found that there is no way to get enough scores for this system to pass it through the University of Tokyo. Entrance exam."
Wu Xiaoru, president of the board of directors of the University of Science and Technology, believes that while the technology is constantly improving, the college entrance examination proposition is also constantly improving. "Now the topics of the college entrance examination are more and more flexible and more comprehensive, so the handling of these problems is becoming more and more difficult for the machine. We are now also communicating with experts from some testing institutions and found that they are on this topic. In the process, many topics must contain a lot of understandings of you. You have to do a lot of logical transformations in the middle to make a problem well."
However, intelligent AI is rapidly improving its ability to learn in the liberal arts. Earlier, Microsoft Xiao Bing received a favorable review on the poem "Sunshine Lost Glass Window" published on May 19, 2017. It is considered to have made great progress.
Will AI subvert the education industry?
Wu Xiaoru introduced the media. At present, the intelligent scoring products of Keda Xunfei can already achieve subjective questions, and even an essay question on Chinese and English can achieve automatic scoring. Will AI subvert the education industry?
Some people believe that artificial intelligence will become the most important solution for the transformation of the education industry in the next 5-10 years. Promoting intelligent education and building a learner-centered teaching model can improve the shortcomings of traditional education and make up for the shortcomings of traditional education. board.
However, Chen Ruifeng believes that AI is not a substitute for a teacher. Because the robot can't explain it instead of the teacher. For example, why is this question correct, why is it wrong, and why the problem-solving ideas are so wrong, why is it correct? However, the AI ​​can tell the teacher what knowledge points of the student have problems and where they should spend more time training, rather than wasting time on the knowledge the student already has.
Zhang Kailei said that during the middle school period, there are about 3,529 test sites, and each of them can do 3-4 questions. It is enough to write 10,000 questions, but most of the students currently have exercises ranging from 30,000 to 40,000. 3/4 of the questions are not required to be written, and 75% of the precious time is wasted.
In the traditional mode, it takes 2-3 days from the student to write the assignment to the teacher's correction, and then to the student's correction. It is possible to change the day after the introduction of the AI.
According to relevant media reports, the first product developed by the prospective cloud science "quasi-star intelligent evaluation system", the front end is a smart pen, the back end is an artificial intelligence system. At the same time that the student answers the question with this pen, the system will collect the answer data for the whole process and transfer it to the cloud with one key. The background system will make a portrait of the student's learning ability. This kind of intelligent evaluation method can help students find out the knowledge defects in time and prompt for targeted tutoring.
Zhang Kailei said in an interview with the media group that achieving personalized learning and greatly improving learning efficiency and achievements is the original intention of Xuebajun to develop intelligent education robots. "China's 9.8 million teachers are too different. It is too difficult for 9.8 million teachers to be uniform. Is it possible to transplant these capabilities to the system level and let the system complete 70% or 80% of the teacher's behavior? ?"
AI cross-domain capabilities are not strong enough
In addition to the math test, Go is another high ground occupied by artificial intelligence. The top of the international Go Che Jie uses "Go God" to describe the artificial intelligence of Google (microblogging). Alphago's performance of winning 3:0 declared that he became the highest level of Go on the planet.
From the birth of the world to the nine-stage certification issued by the Chinese chess house, Alphago took only seven years.
Although Aidam failed to defeat the former college entrance examination mathematics champion, it was astounding to hand over 134 points in 10 points.
In the vertical segmentation field, AI has shown its ability to be underestimated. It is foreseeable that a variety of Alphago will emerge in the future. Some of the jobs in Hisense’s factory use robots instead of labor, which increases efficiency while reducing costs. Customer service business in many industries has been introduced, based on deep learning, natural language understanding and emotional and emotional recognition technology.
However, the current AI can only be unique in the segmentation field. Li Kaifu (microblogging), the founder of the innovation factory, said in an interview with the media that cross-domain, profound and in-depth thinking content, artificial intelligence can not be reached in the next decade. .
Many startups regard artificial intelligence as the next slogan. It seems that no matter what project they are engaged in, they can attract investment by adding AI.
Wang Sheng, a partner of China Merchants Zhiyuan, pointed out that artificial intelligence actually happened decades ago. In recent years, it has been heated up by some investment institutions. However, artificial intelligence is a big trend. There may be some heat fluctuations in the short term, but the potential for the next ten to twenty years will be huge, and the application scenarios will be more and more.
According to iResearch, the global artificial intelligence market will reach 119 billion yuan in 2020, with a compound annual growth rate of 19.7%. In the same period, China's artificial intelligence market will reach 9.1 billion yuan, with a compound annual growth rate of over 50%.
Zhou Shaojun, a partner of Oriental Fuhai, believes that investment institutions will not cast concepts, but will cast a landing scene. Future technologies must be grounded, including service robots, autonomous driving, drones, etc., all of which are scenarios where artificial intelligence is likely to land.
Disclaimer: The electronic reprinted works of E-Commerce Network are as far as possible to indicate the source, and all rights of the owner of the work are not transferred due to the reprint of this site. If the author does not agree to reprint, please inform the site to delete or correct it. Reprinted works may be subject to change in title or content.
Jiangsu Stark New Energy was founded in 2018. It is an emerging new energy manufacturer and trader. We produce and sell high-quality lithium battery cells, such as 3.2V50ah 3.2V100AH 3.2V200AH and other square aluminum-shell lithium iron phosphate batteries. , 18650 3.7V2000mah 3000mah ternary lithium battery cells, etc.
Batteries are mainly used in energy storage lithium battery packs, electric bicycles, electric golf carts, power tools, toys, etc. Our lithium iron phosphate batteries are set to cycle cycles to 6000 times, and battery voltage life in ternary reaches 2000 times ,According to the requirements of charging and discharging in different fields. The discharge rate of our battery design is 1C 2C 3C 5C to provide options
Lithium Ion Battery Cell,3.2V150Ah Lifepo4 Cell,Prismatic Lithium Iron Battery,Lithium Iron Phosphate Battery Lifepo4
Jiangsu Stark New Energy Co.,Ltd , https://www.stark-newenergy.com