Filtering Noisy Dialogue Corpora by Connectivity and Content Relatedness

TL;DR

This paper introduces a scoring method based on connectivity and content relatedness to filter noisy dialogue data, improving neural dialogue agent training quality.

Contribution

It proposes a novel scoring approach for dialogue pair quality, grounded in linguistic research, and demonstrates its effectiveness in filtering datasets for better dialogue models.

Findings

Filtered data improves dialogue response quality

Scoring correlates well with human judgment

Method effectively reduces noise in large datasets

Abstract

Large-scale dialogue datasets have recently become available for training neural dialogue agents. However, these datasets have been reported to contain a non-negligible number of unacceptable utterance pairs. In this paper, we propose a method for scoring the quality of utterance pairs in terms of their connectivity and relatedness. The proposed scoring method is designed based on findings widely shared in the dialogue and linguistics research communities. We demonstrate that it has a relatively good correlation with the human judgment of dialogue quality. Furthermore, the method is applied to filter out potentially unacceptable utterance pairs from a large-scale noisy dialogue corpus to ensure its quality. We experimentally confirm that training data filtered by the proposed method improves the quality of neural dialogue agents in response generation.