A Comparative Study of Controlled Text Generation Systems Using Level-Playing-Field Evaluation Principles

TL;DR

This paper proposes a standardized evaluation framework for controlled text generation systems, revealing significant discrepancies in reported performances and emphasizing the need for fair, reproducible assessment methods.

Contribution

It introduces a level-playing-field evaluation approach that standardizes outputs and uses shared datasets and metrics for fair comparison of CTG systems.

Findings

Re-evaluation shows many systems perform worse than originally reported.

Standardized evaluation reveals substantial discrepancies in system performance.

Highlights the necessity for reproducible and standardized assessment practices.

Abstract

Background: Many different approaches to controlled text generation (CTG) have been proposed over recent years, but it is difficult to get a clear picture of which approach performs best, because different datasets and evaluation methods are used in each case to assess the control achieved. Objectives: Our aim in the work reported in this paper is to develop an approach to evaluation that enables us to comparatively evaluate different CTG systems in a manner that is both informative and fair to the individual systems. Methods: We use a level-playing-field (LPF) approach to comparative evaluation where we (i) generate and process all system outputs in a standardised way, and (ii) apply a shared set of evaluation methods and datasets, selected based on those currently in use, in order to ensure fair evaluation. Results: When re-evaluated in this way, performance results for a representative set of current CTG systems differ substantially from originally reported results, in most cases for the worse. This highlights the importance of a shared standardised way of assessing controlled generation. Conclusions: The discrepancies revealed by LPF evaluation demonstrate the urgent need for standardised, reproducible evaluation practices in CTG. Our results suggest that without such practices, published performance claims may substantially misrepresent true system capabilities.