Knowledge-based end-to-end memory networks

TL;DR

This paper introduces KB-memN2N, a novel end-to-end memory network that incorporates prior knowledge and handles named entities for goal-oriented dialog systems, showing promising results on specific datasets.

Contribution

The paper proposes a new knowledge-based memory network model that integrates prior task knowledge and improves handling of named entities in dialog systems.

Findings

Effective handling of named entities in goal-oriented dialogs

Improved performance on DSTC6 and dialog bAbI datasets

First step towards knowledge integration in end-to-end dialog models

Abstract

End-to-end dialog systems have become very popular because they hold the promise of learning directly from human to human dialog interaction. Retrieval and Generative methods have been explored in this area with mixed results. A key element that is missing so far, is the incorporation of a-priori knowledge about the task at hand. This knowledge may exist in the form of structured or unstructured information. As a first step towards this direction, we present a novel approach, Knowledge based end-to-end memory networks (KB-memN2N), which allows special handling of named entities for goal-oriented dialog tasks. We present results on two datasets, DSTC6 challenge dataset and dialog bAbI tasks.