Abstract: A recurring challenge of crowdsourcing NLP datasets at scale is that human writers often rely on repetitive patterns when crafting examples, leading to a lack of linguistic diversity. We introduce a novel approach for dataset creation based on worker and AI collaboration, which brings together the generative strength of language models and the evaluative strength of humans. Starting with an existing dataset, MultiNLI for natural language inference (NLI), our approach uses dataset cartography to automatically identify examples that demonstrate challenging reasoning patterns, and instructs GPT-3 to compose new examples with similar patterns. Machine generated examples are then automatically filtered, and finally revised and labeled by human crowdworkers. The resulting dataset, WANLI, consists of 108,079 NLI examples and presents unique empirical strengths over existing NLI datasets. Remarkably, training a model on WANLI instead of MultiNLI (which is 4 times larger) improves performance on seven out-of-domain test sets we consider, including by 11% on HANS and 9% on Adversarial NLI. Moreover, combining MultiNLI with WANLI is more effective than combining it with other NLI augmentation sets. Our results demonstrate the potential of natural language generation techniques to curate NLP datasets of enhanced quality and diversity.
Paper Type: long