An Optimization of Conversational System Based Conceptual Natural Language Understanding

A method for dialogue comprehension based on rules-based in-depth parsing and semantic analysis is presented in the abstract of this paper. A semantic evaluation carried out as part of the EVALDA/MEDIA initiative yielded positive findings. We were able to achieve a score comparable to that of statistical systems without the use of a manually annotated corpus, despite the fact that the conditions of the evaluation made it difficult to use a symbolic processing-based method. Our methodology, which depends on a LTAG parser and two ontologies, gives very strong discourse understanding and offices for communicating with an exchange supervisor and the actual application, and we accept it tends to be utilized in genuine discourse frameworks, despite the fact that the engineering has been intended for this mission, which is solely committed to spoken discourse understanding.