1. Research Area
In this laboratory, the research/development of artificial intelligence related problems is focused on natural language processing. Natural language processing is a technology that computers understand and process human language. Our laboratory focuses on the technology of processing Korean especially in natural language. Based on the technology studied for a long time, we provide the best performance spell checker in Korea. it was installed in Hangul 2018/2020 in cooperation with Hancom. In addition, we participate in the ExoBrain project, a representative national artificial intelligence R&D project, to provide natural language processing technology.
2. Research Overview
- Machine reading
It is a task of natural language processing to enter fingerprints and questions and use them to find the correct answers within a fingerprint.
- Spell Checker(Rule-based, Deep learning based)
It is a task of natural language processing that corrects spelling and grammar errors in documents.
- Natural language processing source technology(Part-of-speech tagging, Parsing, etc):
It is a task of natural language processing to research sentence element analysis technology in order to perform natural language processing with a machine.
- Information Retrieval
It is the task of natural language processing to research the technology for quickly and easily searching for desired information from large volumes of data.
- Korean Wordnet
It is a study to effectively use information on the relationship between words such as synonyms, antonyms, and upper/lower words as a database.
3. Research Achievements
- Order of Cultural Merit (Ministry of Culture and Tourism / 2005)
- New Software Grand Prize(Korean Spell Checker) (Ministry of Information and Communication / 2006)
- Technology transfer of 1.34 billion won over the past three years
- In the past 3 years, 10 domestic patent applications / 2 registrations, industry research expenses of 3.23 billion won
- Currently performing 4 tasks including 3 parts of exobrain SW development (Ministry of Science and Technology Information and Communication / 2016 ~ )