Data Augmentation for Neural Online Chat Response Selection

TL;DR

This paper explores novel data augmentation techniques, permutation and flipping, for neural dialog response selection, demonstrating significant improvements in recall metrics across multiple datasets and languages.

Contribution

It introduces a combined data augmentation approach that merges original and synthesized data, enhancing model generalization in dialog response tasks.

Findings

Achieved 1-3 recall-at-1 points improvement over baselines.

Effective across Chinese and English datasets.

Works for both full-scale and small-scale models.

Abstract

Data augmentation seeks to manipulate the available data for training to improve the generalization ability of models. We investigate two data augmentation proxies, permutation and flipping, for neural dialog response selection task on various models over multiple datasets, including both Chinese and English languages. Different from standard data augmentation techniques, our method combines the original and synthesized data for prediction. Empirical results show that our approach can gain 1 to 3 recall-at-1 points over baseline models in both full-scale and small-scale settings.