Towards the Rapid Development of a Natural Language Understanding Module

TL;DR

This paper presents a rapid, learning-based approach for developing natural language understanding modules for conversational agents, enabling non-experts to quickly create effective language models for specific domains.

Contribution

It introduces a classification-based method that simplifies NLU development, demonstrated through art domain question answering and a cinema database interface.

Findings

Effective for domain-specific NLU tasks

Allows non-experts to develop language understanding modules

Facilitates rapid prototyping of conversational agents

Abstract

When developing a conversational agent, there is often an urgent need to have a prototype available in order to test the application with real users. A Wizard of Oz is a possibility, but sometimes the agent should be simply deployed in the environment where it will be used. Here, the agent should be able to capture as many interactions as possible and to understand how people react to failure. In this paper, we focus on the rapid development of a natural language understanding module by non experts. Our approach follows the learning paradigm and sees the process of understanding natural language as a classification problem. We test our module with a conversational agent that answers questions in the art domain. Moreover, we show how our approach can be used by a natural language interface to a cinema database.