- Feb 4, 2013
In April 2012, Research Center for Knowledge Media and Content Science was established in NII, investigating analysis and synthesis of knowledge based on deep understanding of scholarly information, with a particular focus on textual media.
Improvement of information technology enabled us to use computers to analyse enormous volumes of text. However, information is not always represented in a form that we can make use of as it is. For example, in an exam, you face questions which are based on the textbook, but are represented in a different way. When you are writing a new thesis, it is not sufficient to only repeat the words of other researchers of the same field; you should summarize their work in your own words, and also build upon their conclusions, creating new and novel scientific content. Even just focusing on textual media, it is easy to find a number of tasks that are difficult to solve without achieving deep understanding of the content.
This fact awakens us to the importance of the role our intelligence plays when we absorb knowledge from an external source: transforming information to a form that we can use. Current applications of information processing, such as information retrieval systems, mostly depend on surface clues, which are insufficient for deep understanding; in other words, those applications can’t imitate human intelligence.
In KMCS, we are engaged in both theoretical study and task-based empirical study about knowledge extraction by computational text analysis and logical inference from the acquired knowledge. The outcome of these studies will contribute to the development of intelligent applications which can generalize, integrate or infer knowledge by themselves.