As Hori¹⁾ have claimed, a useful structure of information pieces in one context is no more than one section of a phenomenon, and the structure does not work in another context any more. Therefore, the process of knowledge creation is inevitably dynamic and cyclic. The Knowledge Liquidization & Crystallization is a model of the process of knowledge creation developed based on this dynamic aspect of knowledge.

Suppose you are writing a paper, say, to submit to a conference. You refer to a number of books, papers, etc., and your memory in your mind. You do not use such external and internal information pieces as they are, but you excerpt some relevant potions of information to your current writing and fuse them into an idea. This is the process of knowledge creation in practice that KLC describes. What I am going to do is to support this process by implementing a system Knowledge Nebula Crystallizer that conducts KLC.

Knowledge Liquidization

Dissolving knowledge into small granularity that have a core grounded in the real world and that preserve the local semantic relationships around the core

Knowledge Crystallization

Restructuring the relationships among the granular units in accordance with a current context