Dialogue Response Selection with Hierarchical Curriculum Learning

TL;DR

This paper introduces a hierarchical curriculum learning framework for dialogue response selection, which trains models in an easy-to-difficult manner using corpus-level and instance-level curricula, leading to improved performance.

Contribution

It proposes a novel hierarchical curriculum learning approach with two curricula to enhance dialogue response matching models, addressing limitations of random negative sampling.

Findings

Significant performance improvements on three benchmark datasets.

Effective in training various state-of-the-art matching models.

Demonstrates the benefit of curriculum learning in dialogue systems.

Abstract

We study the learning of a matching model for dialogue response selection. Motivated by the recent finding that models trained with random negative samples are not ideal in real-world scenarios, we propose a hierarchical curriculum learning framework that trains the matching model in an "easy-to-difficult" scheme. Our learning framework consists of two complementary curricula: (1) corpus-level curriculum (CC); and (2) instance-level curriculum (IC). In CC, the model gradually increases its ability in finding the matching clues between the dialogue context and a response candidate. As for IC, it progressively strengthens the model's ability in identifying the mismatching information between the dialogue context and a response candidate. Empirical studies on three benchmark datasets with three state-of-the-art matching models demonstrate that the proposed learning framework significantly improves the model performance across various evaluation metrics.