News

• To lead development and application of AI technology in edtech industry via Stella
• AI based test prep services will help students overcome their weaknesses and forecast test questions
• Expected to reduce educational disparity, enabling ST Unitas to become one step closer to achieving its vision, “Creating a world where dreams are realized even for those who are economically disadvantaged.”

(January 11^th, 2018)

ST Unitas, the leading edtech company in South Korea, proposes a new paradigm for the education industry in the fourth industrial revolution by launching AI-powered education services.   

On January 11^th, ST Unitas announced its plans to introduce “Stella”, an education service which identifies and targets weaknesses of each student and predicts test questions by using artificial intelligence (AI). ST Unitas has already led the education market with its “Dangi (Short Term) High Score Methodology”, a study method developed by analyzing big data of high-scoring students collected for the past few years. The company now plans to enjoy an unrivaled position in the edtech industry with the new AI service.

Stella, an AI-powered service developed by ST Unitas, largely provides two types of services. “Weakness Target Service” aims at improving weak points of each individual by analyzing the level of the student and questions that the student frequently gets wrong, and “Test Question Prediction” offers students a collection of questions which are highly likely to be in the test. The company will first launch “Weakness Target Service” in Gongdangi, a test prep brand for civil servant exams, and “Test Question Prediction” in English Dangi, respectively, to pilot test the services before gradually expanding them to other brands across various fields of education. 

In order to enable Stella to provide “Weakness Target Service”, deep learning was used for big data analytics including approximately 300,000 registered users’ study history and 60,000 questions from civil servant tests. The service will also analyze the student’s study status and build a Bayesian Network to realize algorithms optimized for each individual student, which will support the student by predicting which questions the student will get wrong and enable the student focus on such type of questions.

In the process of analyzing the student’s study status, Stella will also collect and organize the questions which the student gave wrong answers to. As a result, ST Unitas is expecting that Stella will greatly reduce not only the time invested in preparing for civil servant exams but also the vast amount of contents students have to study.

Bonjun Koo, who is in charge of AI development for Stella service, stated, “When students study multiple books it’s highly likely that the will have to look through contents that they already know or answer easy questions which aren’t much help to them. Stella excludes such contents and questions and therefore enhances the student’s efficiency in studying for exams. The more learning data accumulates in Stella, the more precise and elaborate Stella’s forecasting service will become by using deep learning.” 

To provide “Test Question Prediction” service, Stella perceives and learns TOEIC data on its own, and automatically extracts question types which are most likely to be on the actual exam. Deep learning of big data, mainly composed of tens of thousands of TOEIC questions which have been gathered and created by English Dangi as well as analysis of actual questions on TOEIC by ST Unitas’ language institute members, allows Stella to automatically process such data. The new service enables students to reduce the amount of time invested in studying for the exam as well as the enormous amount of contents they have to study.

Sunghyuk Yoon, the CEO of ST Unitas, remarked, “The educational environment in the fourth industrial revolution is expected to experience a drastic transformation with the help of AI technology. As ST Unitas foresaw that AI will resolve the educational disparity resulting from income gap as well as time and regional constraints, the company had been preparing the development of Stella. Based on continuous investment and development, we are planning to further develop Stella as the dominant leader in the era of AI education so that in the future education opportunities are evenly provided to everyone throughout the world. We believe in contributing to creating a better world where all people are able to make their dreams come true through education.”

The new AI-powered service, Stella, is available on CONECTS (http://www.conects.com), a knowledge sharing platform of ST Unitas.