Zero-Resource Knowledge-Grounded Dialogue Generation

TL;DR

This paper introduces a zero-resource approach for knowledge-grounded dialogue generation, enabling models to learn from separate dialogue and knowledge data without requiring paired training examples.

Contribution

It proposes a variational method that models knowledge as latent variables, allowing training without knowledge-grounded dialogue triples.

Findings

Achieves comparable performance to state-of-the-art methods.

Demonstrates strong generalization across topics and datasets.

Operates effectively without knowledge-grounded dialogue training data.

Abstract

While neural conversation models have shown great potentials towards generating informative and engaging responses via introducing external knowledge, learning such a model often requires knowledge-grounded dialogues that are difficult to obtain. To overcome the data challenge and reduce the cost of building a knowledge-grounded dialogue system, we explore the problem under a zero-resource setting by assuming no context-knowledge-response triples are needed for training. To this end, we propose representing the knowledge that bridges a context and a response and the way that the knowledge is expressed as latent variables, and devise a variational approach that can effectively estimate a generation model from a dialogue corpus and a knowledge corpus that are independent with each other. Evaluation results on three benchmarks of knowledge-grounded dialogue generation indicate that our model can achieve comparable performance with state-of-the-art methods that rely on knowledge-grounded dialogues for training, and exhibits a good generalization ability over different topics and different datasets.