Graduate School of Knowledge Service Engineering
Department of Industrial & Systems Engineering
Korea Advanced Institute of Science and Technology (KAIST)
Local Arrangement Chair:
7th Korea-Japan Database Workshop (KJDB 2012)
Information Director (Mar. 2007 ~ Aug. 2008):
ACM Trans. on Knowledge Discovery from Data (TKDD)
KIISE Database Society of Korea (Aug. 2011 ~ present)
Korea Spatial Information Society (June 2015 ~ present)
ACM Computing Surveys, The VLDB Journal, IEEE Trans. on Knowledge and Data Engineering (TKDE), Data Mining and Knowledge Discovery (DMKD), Information Sciences, World Wide Web Journal
IoT Group, Samsung Research (May 2017 ~ present)
Big Data Group, Samsung Software Center (Mar. 2013 ~ Feb. 2015)
At KAIST (after professor appointment)
See Projects here
Participation at IBM Research
IBM, "IBM Smart Analytics Optimizer," Sept. 2008 ~ Nov. 2010.
Data warehousing is commonly used by companies to analyze large sets of data for enterprise purposes. Conventional database systems such as IBM DB2 have severe limitations in achieving high analytic performance due to the unprecedented amount of data and high complexity of queries. IBM Smart Analytics Optimizer (ISAO) is a new IBM product for data warehousing to boost analytic performance. It exploits many cutting-edge technologies including in-memory techniques, cache-conscious algorithms, compression, and parallel evaluation of query predicates. Since I joined Dr. Guy Lohman's research team at the early stage of the development, I have made fundamental contributions to the query engine, especially in run-time join, short-circuit evaluation, data types, and many other important aspects. ISAO is proven to significantly outperform IBM DB2 for real customer workloads. ISAO was released to the market as an IBM product in November 2010. Contributing to and releasing an official product is one of the greatest honors at IBM.
Participation at U of I
National Science Foundation (NSF), "MotionEye: Querying and Mining Large Datasets of Moving Objects," July 2006 ~ Aug. 2008.
The design and development of powerful mechanisms for managing and mining large datasets of moving-object information is an emerging direction in science and engineering informatics. The subject becomes increasingly important as the world becomes more and more mobile. Data mining for moving-object trajectories is much more difficult than that for static datasets since trajectories can represent very long and complex movement histories. I have established a new paradigm of trajectory mining that enables us to extract knowledge of partial trajectories. Based on the paradigm, I developed innovative methods for finding typical characteristics of moving-object trajectories, for uncovering suspicious motions in large datasets of moving objects, and for classifying moving-object trajectories. These original research results have all been published in top conferences or journals.
National Science Foundation (NSF), "Movebank for Animal Tracking and Photo Monitoring Data," Feb. 2008 ~ Aug. 2008. Movebank is an interdisciplinary project between animal science and computer science. Thousands of biologists collect animal-movement data, but since there is no tool to save or compare them, most data are used once and then disappear into a filing cabinet. Movebank facilitates services of useful animal-movement resources and long-term comparisons of these data. I developed online mapping and analysis tools for Web service. Especially, our trajectory clustering software has been made available to public via Web service.
Participation at KAIST (during my Ph.D. course)
Korea Science and Engineering Foundation (KOSEF) / Advanced Information Technology Research Center (AITrc), "Distributed Multi-DBMSs (ODYSSEUS)," July 1999 ~ June 2006.
In this project, we were developing the object-relational DBMS ODYSSEUS and extending it to a distributed multi-DBMS. I was a chief architect of ODYSSEUS for the years 2001 through 2003. In particular, I implemented the object view, access control mechanism, and user-defined type as well as improved core part of the query processing module for tightly coupling the DBMS with information retrieval (IR) features. Our work for tight-coupling with IR features received the Best Demonstration Award from the prestigious IEEE ICDE conference.
Ministry of Science and Technology, "Development of Web-DBMS Integration Technology for a Web-Based OODBMS Tightly-Coupled with Information Retrieval," Dec. 1998 ~ Sept. 2000.
I developed the web-DBMS gateway ODYSSEUS/Web for the ODYSSEUS ORDBMS. The gateway was a very novel and advanced solution at that time in that it was fully integrated into web servers. ODYSSEUS/Web simplified the development of web-based applications by using only HTML tags. This accomplishment significantly improved the usability of the ODYSSEUS ORDBMS.
MySeek Co., "Engineering of ODYSSEUS for the MySeek Web Search Engine," June 2000 ~ May 2001.
We developed a web search engine using the ODYSSEUS ORDBMS. I was a chief architect of the search engine, which provided nationwide services and had at-that-time innovative features such as Home Search.