6533b831fe1ef96bd129831e

RESEARCH PRODUCT

Dialog Structure Automatic Modeling

Débora HisgenDaniela López De Luise

subject

Structure (mathematical logic)Test caseComputer sciencebusiness.industryStatistical analysisArtificial intelligenceDialog boxArchitecturecomputer.software_genrebusinesscomputerWord (computer architecture)Natural language processing

description

This paper presents the approach implemented as part of a conversational bot named WIH (Word Intelligent Handler). It has a complex architecture with several components. Some of them are the ER-memory, EP-memory and other minor modules that provide the prototype of good modeling of Spanish sentences. They constitute the knowledge-representation mechanism that is used by WIH to build automatic answers during dialogs with humans. In this paper a small description of these components and some of its interaction is given, along with test cases and a statistical analysis of the results obtained. It is shown here that WIH prototype can adapt its behavior and learning rate of its internal working memories according to dialogs contents.

yearjournalcountryeditionlanguage
2010-01-01
https://doi.org/10.1007/978-3-642-16761-4_7