Towards an Automatic Turing Test: Learning to Evaluate Dialogue Responses
This addresses a crucial need in dialogue research for accurate automatic evaluation to reduce reliance on expensive human assessments, though it is incremental as it builds on learning-based approaches.
The paper tackles the problem of automatically evaluating dialogue response quality, where existing metrics poorly correlate with human judgments, by introducing ADEM, a model that learns to predict human-like scores and shows significantly higher correlation with human judgments than metrics like BLEU.
Automatically evaluating the quality of dialogue responses for unstructured domains is a challenging problem. Unfortunately, existing automatic evaluation metrics are biased and correlate very poorly with human judgements of response quality. Yet having an accurate automatic evaluation procedure is crucial for dialogue research, as it allows rapid prototyping and testing of new models with fewer expensive human evaluations. In response to this challenge, we formulate automatic dialogue evaluation as a learning problem. We present an evaluation model (ADEM) that learns to predict human-like scores to input responses, using a new dataset of human response scores. We show that the ADEM model's predictions correlate significantly, and at a level much higher than word-overlap metrics such as BLEU, with human judgements at both the utterance and system-level. We also show that ADEM can generalize to evaluating dialogue models unseen during training, an important step for automatic dialogue evaluation.