Md.K. Rhaman and T. Endo (Japan)
Three Layer Conceptual Network and knowledge representation.
Contextual analysis in dialog is still a hard problem. In this paper three layers memory structure is adopted to address the challenge which we refer to as “Three Layer Conceptual Network” (TLCN). This highly efficient network simulates the human brain by episodic memory, discourse memory and ground memory. It can understand ambiguous incoming utterances and generate responses by using the accumulated discourse information. Knowledge representation is one of the big challenges for Natural Language Processing. We use an extended case structure framework to represent the knowledge. The knowledge database is constructed by the collection of target system information and utterances. This knowledge updates after every dialog conversation. A classifier is also introduced for classifying the knowledge for the target system. This system prototype is based on doctor patients dialogs. 63% disease identifying accuracy is observed by this system prototype.
Important Links:
Go Back
