Synthesizing Adversarial Negative Responses for Robust Response Ranking and Evaluation

TL;DR

This paper introduces data-driven methods to generate adversarial negative responses that improve the robustness of dialogue response ranking and evaluation models by focusing on coherence and appropriateness beyond mere content similarity.

Contribution

It proposes mask-and-fill and keyword-guided techniques for creating challenging negative examples, enhancing model training for more coherent and contextually appropriate responses.

Findings

Outperforms baseline methods in response ranking accuracy.

Generates adversarial responses that are contextually similar but incoherent or inappropriate.

Improves robustness of dialogue models across multiple datasets.

Abstract

Open-domain neural dialogue models have achieved high performance in response ranking and evaluation tasks. These tasks are formulated as a binary classification of responses given in a dialogue context, and models generally learn to make predictions based on context-response content similarity. However, over-reliance on content similarity makes the models less sensitive to the presence of inconsistencies, incorrect time expressions and other factors important for response appropriateness and coherence. We propose approaches for automatically creating adversarial negative training data to help ranking and evaluation models learn features beyond content similarity. We propose mask-and-fill and keyword-guided approaches that generate negative examples for training more robust dialogue systems. These generated adversarial responses have high content similarity with the contexts but are either incoherent, inappropriate or not fluent. Our approaches are fully data-driven and can be easily incorporated in existing models and datasets. Experiments on classification, ranking and evaluation tasks across multiple datasets demonstrate that our approaches outperform strong baselines in providing informative negative examples for training dialogue systems.