Show Your Work: Improved Reporting of Experimental Results

TL;DR

This paper highlights the importance of reporting validation performance during model development, introduces a method to estimate expected validation performance based on computation, and advocates for improved reporting practices in NLP research.

Contribution

It proposes a novel technique to estimate expected validation performance as a function of computation, enhancing the robustness of model comparisons.

Findings

Validation performance reporting varies widely across papers.

Estimated computation time to reach certain accuracy levels varies from hours to weeks.

Using the proposed method can change conclusions about model superiority.

Abstract

Research in natural language processing proceeds, in part, by demonstrating that new models achieve superior performance (e.g., accuracy) on held-out test data, compared to previous results. In this paper, we demonstrate that test-set performance scores alone are insufficient for drawing accurate conclusions about which model performs best. We argue for reporting additional details, especially performance on validation data obtained during model development. We present a novel technique for doing so: expected validation performance of the best-found model as a function of computation budget (i.e., the number of hyperparameter search trials or the overall training time). Using our approach, we find multiple recent model comparisons where authors would have reached a different conclusion if they had used more (or less) computation. Our approach also allows us to estimate the amount of computation required to obtain a given accuracy; applying it to several recently published results yields massive variation across papers, from hours to weeks. We conclude with a set of best practices for reporting experimental results which allow for robust future comparisons, and provide code to allow researchers to use our technique.