Classi Co., Ltd., a joint venture between Benesse Holdings Corp. and SoftBank Corp., is collaborating with Tokyo Institute of Technology to conduct empirical research to put into practical use a mechanism for automatically creating problems using natural language processing in 2018. It will start in January.

 "Classi" is a learning support cloud service that connects teachers, students, and parents in school lessons, student guidance, and student learning.It has been introduced to more than 2,100 high schools nationwide and is used by more than 80 people (as of December 2017). "Adaptive learning learning videos" that recommend optimal learning from more than 12 units and 1,000 levels of difficulty, 6 "video teaching materials" and "major publishers 2.5" as learning content according to each student's academic ability It provides a collection of questions such as "A collection of questions."

 "Classi" has already provided learners with about 7 problems, but in the future, as the number of users increases and students of various progress repeat learning, more problems will be needed. Things are expected.At that time, if a self-operation question system is realized, it will be possible to quickly create a large number of various problems, and it will be possible to provide optimal problems according to academic ability more than ever.

 Therefore, in cooperation with the Tokunaga Kenshin Laboratory of Tokyo Institute of Technology, we will start empirical research aiming at the construction and practical application of a self-operation question system.Specifically, it asks the question of the meaning of English words by itself and distributes it on the Classi platform.Students (about 1 first-year high school students) answer the question and evaluate the correlation between the student's academic ability measured in advance by the test and the correct answer rate of the question.In addition, before asking a question to a student, a check by an expert in question creation is performed in advance, and it is carried out after confirming that a valid question is created.

 The flow of the automatic questioning system is as follows.
(XNUMX) The questioner specifies the target word and its meaning from the dictionary prepared in advance.
(XNUMX) Search the text on the Internet that the machine uses with the specified meaning of the target word, and create the question text from there.
③ Use the synonym of the target word as the correct answer using the dictionary used in ①.
④ Select a word that has a different meaning from the target word from the question text or dictionary, and create an incorrect answer option.
⑤ The difficulty level of the question is adjusted by the difficulty level of the question text, the closeness of the meaning of the wrong answer and the correct answer, and the difficulty of the word of the wrong answer itself.

 In addition to this kind of learning support, "Classi" is also important in school activities, such as "school group information" that is the basis of communication such as school, grade, class, club activities, and various things in the school. We are also supporting the introduction of ICT in all school activities through "student charts" that can accumulate activity results.

Source: [Classi Co., Ltd.] Classi, Tokyo Institute of Technology and Tokyo Institute of Technology start self-operation question empirical research on machines using natural language processing

Tokyo Institute of Technology

The pinnacle of science and engineering university that continues to produce science and engineering people with high aspirations and a Japanese spirit, mastering the knowledge that creates the times, refining their skills

Tokyo Institute of Technology was established as the Tokyo Institute of Technology in 1881 (Meiji 14), when modernization of industry was an urgent need.Since its establishment, it has continued to produce excellent research results with excellent science and engineering human resources, and is still at the top of Japan's science and engineering universities.Tokyo Institute of Technology requires not only a high degree of specialization but also liberal arts […]

University Journal Online Editorial Department

This is the online editorial department of the university journal.
Articles are written by editorial staff who have a high level of knowledge and interest in universities and education.