High Performance in Text Processing: the CPP-TRS Environment

Graziella Tonfoni

A dynamic systenl for integrating natural language and vision is presented, specifically aimed toward supporting an interpretative model and a set of conceptual tools, which may consistently represent those qualitatively different aspects of natural language, which are opaque and therefore difficult to identify. Qualitative reasoning about conmmnication via natural language is being proposed as to allow identification and consistent definition of different values of each conununicative action performed and reflecting upon texts, text segn~ents and text maits. A visual representation system and a highly articulated and precise terminolo~' are being used in order to explicitate results of observation of close relationships existing between language and visual perception phenomena as well as relevant concretions. Ternfinology coming from vision is also being redefined as to adapt to precise definition of very specific natural language problems at various levels of complexity such as ambiguity, context dependency and non systematicity. CPP-TRS means precisely Communicative Positiomng Program-Text Representation Systems, the main concept the new framework is based upon is that text processing can be effectively enhanced first recognizing various and qualitatively different kinds of knowledge representation processes, which are involved and then visualizing them consistently. Based on a very simple syntax. CPP-TRS is aimed toward representing and conveying comnmnicative function, intention and turn-taking vistmlly.

This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.