Relations World: A Possibilistic Graphical Model

TL;DR

This paper introduces a possibilistic graphical model for dialog systems, demonstrating its ability to infer family relations, recover from errors, and adapt to context, with theoretical insights into the model's foundations.

Contribution

It presents a novel possibilistic graphical model for world modeling in dialog systems, focusing on relation inference and error recovery.

Findings

Successfully derived family relation models from dialog

Improved error recovery from coreference ambiguities

Explored theoretical properties of the graphical model

Abstract

We explore the idea of using a "possibilistic graphical model" as the basis for a world model that drives a dialog system. As a first step we have developed a system that uses text-based dialog to derive a model of the user's family relations. The system leverages its world model to infer relational triples, to learn to recover from upstream coreference resolution errors and ambiguities, and to learn context-dependent paraphrase models. We also explore some theoretical aspects of the underlying graphical model.