Personal Computer: My New English Teacher
Dialogue-based English Language Learning Service
GenieTutor, a type of dialogue-based English language learning technology, is one of ETRI’s best known research outcomes. Many people have difficulty communicating in English at a time when the language is increasingly becoming an essential tool for global communication. This technology, however, offers a fast, easy way to learn foreign languages. In this regard, GenieTutor was selected as ETRI’s top achievement and as one of the top ten achievements of government-funded research institutes in 2013. As one of the top 100 achievements of national research and development in 2014, ETRI’s dialogue-based English language learning technology has been acclaimed both domestically and internationally. Created to help all Koreans communicate freely in English, GenieTutor combines the IT and education sectors to present a more advanced, convergent service and demonstrate ETRI’s endeavors for better technology.
01
Can’t we learn English
without English-speaking teachers
With a growing demand for English-speaking ability in daily life, people are pressed to find efficient ways to improve their language proficiency. Developing fluency in the English language, the lingua franca of today, is becoming important against the backdrop of the global market’s openness and resulting market integration. However, being familiar with the language is not a simple task.
In Korea, this has become a common social issue, placing a significant economic burden on households due to the excessive enthusiasm over private English tutoring. Since tutoring costs soar when a native English speaker is hired for conversation courses, it has come to be urgently required to develop an alternative technology. Further, as Korea witnesses increasingly more multicultural families, addressing their communication issues is now regarded an important national responsibility.
In this context, GenieTutor, the foreign-language voice conversation software developed by ETRI, has arrived as a highly innovative technology. It is anticipated not only to help reduce foreign-language private education expenses, amounting almost KRW 15 trillion each year, but also to contribute to facilitating the communication ability of multicultural families.
02
What if we couldpractice
Englishwith our computer
As noted, the global language-learning market has been growing constantly, amounting to USD 58.2 billion in 2011. The market revenue of integral voice recognition technology reached USD 46.9 billion in 2011 and is expected to grow to USD 113.2 billion in 2017, according to expert forecasts. To respond to this market demand and to help address this unexpected social issue in Korea, ETRI in 2013 developed the world’s first software that allows personal computers to tutor and correct English expressions, grammar, and pronunciation.
Called GenieTutor, this dialogue-based English language learning service is the outcome of a project sponsored by the Ministry of Science, ICT and Future Planning to develop core technologies in the software computing industry. Allowing users to practice English listening and speaking while talking with their computers in English, the program extends voice recognition and language processing technologies. The world has a quite long history of voice recognition, language processing, and automatic translation technologies. Researchers around the world have studied related topics for decades, and ETRI launched its research on those technologies in the 1980s. Although some large companies have joined in the endeavor, few have continued their research toward highly advanced technologies. Founded as a government-funded institution, ETRI, based on a sense of duty, has never ceased to invest in developing advanced technologies.
03
Program to correctEnglish expressions, grammar,and pronunciation
GenieTutor’s core technology is the natural-language conversation interface. As the name suggests, this technology enables computers to recognize the language used in daily lives and use the data for mutual communication; it was derived from voice recognition technology, which has come into the limelight in recent years. Back in 2013, however, commercialization of voice recognition was in its infancy, because it was never easy to program computers to independently recognize the natural language uttered by different people and respond to them with appropriate sentences. In addition, at the time that ETRI tried to apply the technology to a program designed for foreign-language education, a significantly high level of technology was required.
The history of GenieTutor dates back to 2006, when voice recognition technology was developed. The industrial sectors for this new growth engine at the time included technologies for telematics and home automation. While engaging in those areas, ETRI began to develop voice recognition technology for robots and successfully developed, in 2008, a navigation system that could recognize verbal input for destinations. Developed on the foundation of this type of voice recognition technology, GenieTutor went beyond the simple voice recognition and developed into a conversational interface that can understand what users say and offer suitable responses to them. ETRI also calibrated the voice recognition technology to best serve the English pronunciation used by Koreans, allowing computers to correct grammar, pronunciation, or expression errors frequently made by Korean users.
04
Helping communication of multicultural families and facilitating the Korean wave
A total of 30 core technologies comprising GenieTutor have been transferred; these are worth over KRW 2.55 billion or 14.8% of the research funds invested thus far. Considering that the average ratio of technology transfer at the world’s top research institutes corresponds to about 8% of their research funds, ETRI’s achievement comes as a highly satisfactory result. Also widely noted is the fact that GenieTutor is the world’s first English language-learning service system based on a conversational voice interface. As similar existing services can perform only simple pronunciation evaluation using voice recognition, GenieTutor is praised as an innovative program.
In addition, ETRI solved the problem of existing technologies: challenges to commercializing the correction feedback and dialogue processing technologies because of their high level of difficulty. As a result of ETRI’s analysis of the economic feasibility, GenieTutor is expected to bring an economic ripple effect of KRW 755.5 billion in the sales connected to R&D efforts and to create 4,933 jobs by 2020. Currently, ETRI is conducting research to perfect the core technologies of GenieTutor and expand the service’s application.
In addition to providing foreign-language learning services to Korean users, ETRI also aims to reach more users by developing a service to teach the Korean language to foreigners. This reflects the institute’s commitment to addressing the communication challenges faced by multicultural families and strengthening the status of the Korean wave. In this regard, as part of its export-oriented plans, ETRI developed technologies customized for the vocalization of foreign users in China, Japan, and Southeast Asia.
As mentioned earlier, GenieTutor is a combination of various technologies, including voice recognition, dialogue processing, and other artificial intelligence technology specifically targeting language. ETRI researchers expect that they will be able to further upgrade the service when the unnatural elements of the current conversation interface technology are integrated with more advanced technologies such as deep learning.
- 1980.
- Began research on voice recognition, language processing, and automatic translation.
- 2008.
- Developed voice-recognition navigation system.
- 2015.
- Developed GenieTutor 1.0, a dialogue-based language learning service.
- 2016-
- Developing GenieTutor 2.0 by upgrading existing technologies.