Towards Automatic Evaluation of Dialog Systems: A Model-Free Off-Policy Evaluation Approach

by   Haoming Jiang, et al.

Reliable automatic evaluation of dialogue systems under an interactive environment has long been overdue. An ideal environment for evaluating dialog systems, also known as the Turing test, needs to involve human interaction, which is usually not affordable for large-scale experiments. Though researchers have attempted to use metrics (e.g., perplexity, BLEU) in language generation tasks or some model-based reinforcement learning methods (e.g., self-play evaluation) for automatic evaluation, these methods only show a very weak correlation with the actual human evaluation in practice. To bridge such a gap, we propose a new framework named ENIGMA for estimating human evaluation scores based on recent advances of off-policy evaluation in reinforcement learning. ENIGMA only requires a handful of pre-collected experience data, and therefore does not involve human interaction with the target policy during the evaluation, making automatic evaluations feasible. More importantly, ENIGMA is model-free and agnostic to the behavior policies for collecting the experience data (see details in Section 2), which significantly alleviates the technical difficulties of modeling complex dialogue environments and human behaviors. Our experiments show that ENIGMA significantly outperforms existing methods in terms of correlation with human evaluation scores.


page 24

page 28

page 32

page 35

page 37


Towards an Automatic Turing Test: Learning to Evaluate Dialogue Responses

Automatically evaluating the quality of dialogue responses for unstructu...

Human vs Automatic Metrics: on the Importance of Correlation Design

This paper discusses two existing approaches to the correlation analysis...

Investigating Evaluation of Open-Domain Dialogue Systems With Human Generated Multiple References

The aim of this paper is to mitigate the shortcomings of automatic evalu...

Evaluating Creativity in Computational Co-Creative Systems

This paper provides a framework for evaluating creativity in co-creative...

How To Evaluate Your Dialogue System: Probe Tasks as an Alternative for Token-level Evaluation Metrics

Though generative dialogue modeling is widely seen as a language modelin...

Model-based Bayesian Reinforcement Learning for Dialogue Management

Reinforcement learning methods are increasingly used to optimise dialogu...

CheerBots: Chatbots toward Empathy and Emotionusing Reinforcement Learning

Apart from the coherence and fluency of responses, an empathetic chatbot...