GENIE: Toward Reproducible and Standardized Human Evaluation for Text Generation
This addresses the need for reproducible and standardized human evaluations in text generation research, which is an incremental improvement over existing practices.
The authors tackled the problem of inconsistent and non-reproducible human evaluation in text generation by developing GENIE, a system that standardizes evaluations across tasks like machine translation and summarization, resulting in a public leaderboard that has already ranked 50 submissions from 10 research groups.
While often assumed a gold standard, effective human evaluation of text generation remains an important, open area for research. We revisit this problem with a focus on producing consistent evaluations that are reproducible -- over time and across different populations. We study this goal in different stages of the human evaluation pipeline. In particular, we consider design choices for the annotation interface used to elicit human judgments and their impact on reproducibility. Furthermore, we develop an automated mechanism for maintaining annotator quality via a probabilistic model that detects and excludes noisy annotators. Putting these lessons together, we introduce GENIE: a system for running standardized human evaluations across different generation tasks. We instantiate GENIE with datasets representing four core challenges in text generation: machine translation, summarization, commonsense reasoning, and machine comprehension. For each task, GENIE offers a leaderboard that automatically crowdsources annotations for submissions, evaluating them along axes such as correctness, conciseness, and fluency. We have made the GENIE leaderboards publicly available, and have already ranked 50 submissions from 10 different research groups. We hope GENIE encourages further progress toward effective, standardized evaluations for text generation.