Artificial intelligence begins to teach your children!
Hearing this sentence, many parents may be a little confused. However, in the Guangzhou Branch of Huanggang Center in Hubei Province, this is a fact that is happening.
Hu Yu, CEO of Iflytek and president of consumer business group, recently revealed that Iflytek selected some students from one grade and two classes in Huanggang Center, Hubei Province, and equipped them with a new "robot teacher". This "teacher" can not only assign and correct homework for students, but also analyze the problems and shortcomings after correction. At present, the level of artificial intelligence teachers is not bad. —— After a semester's experiment, the students' grades with personalized learning recommendation system jumped to the top 30 in the whole grade 16 class.
After business, security, transportation and other industries, education has obviously become the new goal of AI. Can you do it without a teacher in a hundred years?
The machine teacher invaded the classroom.
Whether artificial intelligence can replace teachers has long attracted much attention. According to the survey of more than 2,000 kinds of work automation possibilities initiated by Forbes and many institutions, teachers are the least likely to be replaced by machines.
"But this does not mean that AI has lost its job opportunities in the education industry." Zhang Jian, Ph.D., School of Computer Science and Technology, Beijing University of Posts and Telecommunications, told the Science and Technology Daily reporter, "If you don't have to explore philosophical issues such as the meaning of life and answer basic analytical questions, that is the strength of robots."
At present, the global artificial intelligence industry has carried out relevant practices to a certain extent. For example, Japan's Softbank put its own intelligent robot Pepper in Waseda Zhengchigo, Fukushima Prefecture, Japan to go to school with students to help them learn English and computer knowledge; In a class of more than 300 people at Georgia Institute of Technology, Jill Watson worked as an artificial intelligence robot as a teaching assistant for three months without being discovered by students. 20 17 1, Professor Wang Wei of Beihang University walked into Zhongguancun No.1 Primary School with his artificial intelligence robot "Xiao Pang". Under his leadership, Xiao Pang completed English follow-up, projection and self-built map. ...
In Iflytek's experiment, after correcting the homework, the artificial intelligence teacher can accurately analyze the problem and give the students homework. Hu Yu believes that if every student can have such a wise teacher to give expert training and guidance, his or her grades will obviously improve quickly.
What does the AI teacher do?
At present, the application of artificial intelligence technology in education is mainly reflected in two aspects: image recognition and speech recognition. But for the future, AI teachers can obviously do more.
One is adaptive learning. Using adaptive learning technology, we can collect all kinds of learning data of students, predict their future performance, push the most suitable learning content and improve their learning effect.
The second is content analysis. Establish a model to identify and optimize content, and build a knowledge map, so that users can find their own content more easily and accurately, such as a graded reading platform, and use AI technology to rewrite or push appropriate reading content for students with different reading levels.
The third is intelligent evaluation. How to use artificial intelligence to realize large-scale and personalized homework feedback is an important scene of the combination of artificial intelligence and education. Using speech recognition, semantic recognition, automatic error correction or classification, we can not only realize large-scale automatic error correction, but also give personalized feedback.
Specific to the daily teaching activities, the tasks that AI teachers can take over include automatic correcting homework, online answering questions, speech recognition evaluation, personalized learning and so on.
Among them, the progress of semantic analysis technology makes it possible to automatically correct homework. For a simple literal grammar machine, it can automatically identify and correct mistakes, and even put forward suggestions for modification, which will greatly improve teachers' teaching efficiency. At present, the application of speech recognition technology in education is mainly used for oral English assessment. The software developed by Iflytek, Qingrui Education and 5 1Talk can quickly evaluate pronunciation and point out inaccurate pronunciation during reading, and train users' oral English through repeated evaluation.
In addition, with the support of big data, personalized education may be the greatest value that artificial intelligence brings to education.
Big data can describe the learning characteristics of each student. According to the analysis of a research institute in London, people's learning methods can be divided into 70 kinds; Robots in an institution have accumulated 800 million problems that 6.5438+0.3 million students have done, which provides sufficient basis for personalized teaching.
According to foreign media reports, the famous American information service provider McGraw-Hill Group is developing digital courses. It collects information of 2 million students and uses artificial intelligence to create an adaptive learning experience for each student. When students read the material and answer questions, the system will give relevant information according to the students' mastery of knowledge. The system knows what questions students should take and what methods are more acceptable to students. The system will also save students' information for as long as possible, so as to bring more help to students in the future.
Finally, artificial intelligence can also feedback and evaluate the teaching system. With the help of big data, through the statistics of students' learning growth process and effectiveness, we can diagnose students' differences in knowledge, ability structure and learning needs, thus helping students and teachers to obtain real and effective diagnostic data. Students can see problems clearly and learn more efficiently; Teachers can also choose different teaching objectives and contents according to specific conditions and implement different teaching methods to further improve the pertinence, effectiveness and scientificity of teaching and learning.
The premise of coming to power is big data.
As schools contact the Internet to reduce costs and increase profits, the application of AI in the education industry will be more and more extensive. The report released by NMC/CoSN (American New Media Education Alliance and School Networking Alliance) points out that AI and VR will be the two most important technologies that will change the education industry in the next four to five years. Technavio, an American market research company, said that in the next four years, AI's "penetration" into the education field will continue to strengthen, and its compound annual growth rate will be as high as 39%.
"The essence of artificial intelligence is to surpass ordinary people through learning, reach or approach the level of ordinary experts in knowledge and common sense, and the marginal expansion cost of its replication and use is extremely low." Hu Yu said.
However, insiders pointed out that artificial intelligence cannot completely replace teachers, at least not now. The role of teachers is not only to impart knowledge, but more importantly, to teach students to be human and shape their quality through emotional input and ideological guidance. Even if artificial intelligence surpasses human beings in knowledge reserve, knowledge dissemination speed and teaching methods, human teachers still play an irreplaceable role.
Artificial intelligence technology itself also needs to solve the excessive dependence on big data. Like other industries, big data is the premise of artificial intelligence. Wu Xiao, senior vice president of Iflytek, said earlier that the application of artificial intelligence in education starts with data collection, that is, education big data.
"The hardest part of big data education is data collection. A prerequisite for educating big data is whether we can get data. " Wu Xiaoru said.
He said that a prerequisite for artificial intelligence to improve teaching activities is that teaching activities can be structured, and anything that can't be structured can't be assisted by computers. Then, what links can be structured in teachers' teaching decision-making, whether teachers' teaching behavior can be structured, and whether some key tasks with heavy workload, such as correcting homework papers, can be structured. Whether these data can be used in the future needs a good teaching evaluation method to connect. Only in this way can we form the data needed for future education and theory in the real digital environment, and these data can form various models to guide learning through some high-quality teaching experiences in some teaching models.
"This requires that with the support of regional data assets and data assets of data centers jointly built by schools, the whole teaching activities, examinations and learning can form a virtuous circle and promote students' efficient learning more pertinently." Wu Xiaoru said.
Although the impact of robot peers will not come immediately, human teachers should have a sense of crisis and reform, thinking about how to cultivate irreplaceable ability, what is real education, what kind of talents should be cultivated in future education, and so on. Turn challenges into opportunities, change traditions and innovate education.