Automatic Metrics in Natural Language Generation: A Survey of Current Evaluation Practices

TL;DR

This survey reviews current practices in using automatic metrics for evaluating natural language generation, highlighting issues like misuse and lack of transparency, and offers recommendations for improving evaluation rigor.

Contribution

It provides a comprehensive analysis of how automatic metrics are used in NLG evaluation and identifies key shortcomings and areas for improvement.

Findings

Widespread inappropriate metric usage

Lack of detailed reporting on metric implementation

Missing correlations between automatic metrics and human judgments

Abstract

Automatic metrics are extensively used to evaluate natural language processing systems. However, there has been increasing focus on how they are used and reported by practitioners within the field. In this paper, we have conducted a survey on the use of automatic metrics, focusing particularly on natural language generation (NLG) tasks. We inspect which metrics are used as well as why they are chosen and how their use is reported. Our findings from this survey reveal significant shortcomings, including inappropriate metric usage, lack of implementation details and missing correlations with human judgements. We conclude with recommendations that we believe authors should follow to enable more rigour within the field.