KEYNOTE SPEECH

Korean Semantic Resources & Korean Semantic Processing Systems

Speaker

Cheol-Young Ock


• Professor, School of IT Convergence, University of Ulsan, Korea

Abstract

For semantic processing such as sentence understanding and QA, many sematic resources are needed. One of most important semantic resources is a semantic network. Prof. Ock has been constructing a Korean WordNet, namely Ulsan Word Map (UWordMap) since 2002. The Ulsan Word Map has semantic hierarchy(hypernym and hyponym) like WordNet, but it has another important information, subcategorization of argument of predicate. In this speech Prof. Ock will introduce a structure of the UWordMap and how to use for word sense disambiguation of Korean and for understanding a Query in Exo Brain SW.

Word representation, an important area in natural language processing(NLP) used machine learning, is a method that represents a word not by text but by distinguishable symbol. Existing word embedding employed a large number of corpora to ensure that words that were position near by in text. However, corpus-base word embedding need a lot of corpora because the frequency of word occurrence and increase the number of words. Another word embedding(Sense Vector) is done using dictionary definitions and semantic relationship information(hypernyms and antonyms). Then similar sense’s words have similar vector. Furthermore, it was possible to distinguish vectors of antonym words. In this speech Prof. Ock will introduce the Sense Vector and its applications of sematic operations.

Bio

Cheol-Young Ock is a professor of the school of IT convergence, University of Ulsan, Ulsan, Rep. of Korea. He received his BS (1982), MS (1984), and PhD (1993) degrees in computer engineering from the National University of Seoul, Seoul, Rep. of Korea. He has been a visiting professor at the Russia Tomsk Institute (1994) and Glasgow University (1996). He has been the chairman of the sigHCLT (2007~2008) in the KIISE, Rep. of Korea. He has been a visiting researcher at the National Institute of Korean Language (2008). He received an honorary doctorate from the department of Information and Computer Science, National University of Mongolia, Ulaanbaatar, Mongolia (2007). He received a government medal as a meritorious engineer of Korean language development (2016). He has been constructing a Korean WordNet, namely Ulsan Word Map (UWordMap) since 2002. His research interests include natural language processing (WSD), machine learning, and text mining.

• Home page: http://klplab.ulsan.ac.kr
• E-mail: okcy@ulsan.ac.kr