At the 2020 Artificial Intelligence and Education Big Data Conference held at the end of 2020, Liu Limin, president of the China Education International Exchange Association, said that the education field has become one of the important application scenarios of artificial intelligence.
Constructing a networked, digital, intelligent, and lifelong education system to support and lead the modernization of education is the core of intelligent education and the goal of promoting the deep integration of artificial intelligence and education and teaching.
It is understood that artificial intelligence has been deeply integrated with the education field, and personalized learning using big data has covered more than 10 million students.
In the 2020 college entrance examination paper marking work, 9 provinces have introduced machine marking for subjective propositions such as Chinese and English composition.
However, most experts in the field of education, especially those engaged in composition teaching, believe that subjective topics, especially composition, cannot be simply judged by artificial intelligence. It is not advisable to use a set of digital "standards" to measure the level of composition.
Jia Lijuan, a super teacher who has been teaching composition for more than 30 years, said that the emotions, thoughts, and metaphors contained in the composition are difficult to judge by the current artificial intelligence.
She said: "If artificial intelligence is used to judge composition, there may be students who deliberately use the artificial intelligence judgment system as the standard when writing essays, which loses the meaning of composition assessment."
At present, the artificial intelligence composition correction function is mainly realized by three core modules: scoring module, comment module and error correction module.
Among them, the scoring model framework combines traditional methods and deep learning methods, with words, sentences, and texts as the core scoring dimensions; the comment module provides full-text comments, vocabulary recommendations, and sample essay recommendations; the error correction module provides 14 categories of grammar Error identification and correction, including shallow spelling errors, deep grammatical errors, etc.
It should be said that this is a huge advancement in artificial intelligence technology.
However, in terms of current technical conditions, artificial intelligence judgment composition is currently limited to a certain auxiliary role.
The current practice is to combine machine scoring with manual scoring. When there is a big difference between the results of machine and manual scoring, the machine will automatically select the test paper and submit it to the expert group for final judgment, thereby reducing the difference in subjective evaluation and making subjective The evaluation of the questions is more objective and fair.
In the future, it is very likely that it is not artificial intelligence that will correct composition, but teachers who will use artificial intelligence.
Teachers apply artificial intelligence technology to educational practice, explore student interests, integrate educational resources, and design educational products that meet student needs based on student feedback and data analysis. This may be the deep integration of artificial intelligence and education that people want to see.
"The practice of smart education should focus on summing up experience, making up for shortcomings, promoting the in-depth integration of smart technology and education, giving full play to the advantages of smart technology in the allocation of educational resources, and giving full play to the main role of teachers in the use of smart technology in teaching practice. Provide high-quality education." Liu Limin said.
Reporter Joe Di