Semantics is the study of linguistic meaning. In both computational and non-computational terms, a symbol (e.g. a word) which represents some entity in the real world is replaced by another set of symbols (e.g. dictionary definition, semantic net) and there is a wide consensus that this explains the meaning of the given symbol. However, as the history of Artificial Intelligence shows, there is a wealth of implicit information behind such symbols and linguistic symbols are highly context-sensitive. Representing fully even the simplest concepts (like \"apple\", \"fruit\", etc.) becomes a major task to confront Knowledge Acquisition Problem. In this talk, I will present an overview of a model which utilizes Case-Based Reasoning and looks meaning of a symbol as a situation stored as perceptual information. Case-Based Reasoning is an alternative to standard Rule-Based Systems, where information is kept in pieces of knowledge called cases, and retrieved as necessary per context. These retrieved cases are adapted for new problems using domain knowledge and the results of adaptation is stored for learning. It allows incremental learning where even a small number of cases can be used to make reasoning and learning (much like child language acquisition) is through interaction. These make it suitable to base such a semantic model upon.

DATE: 30 April, 2007, Monday@ 15:40
PLACE: EA 409