Generating Datasets with Pretrained Language Models
This addresses the need for efficient sentence embeddings without human labeling effort, though it is incremental as it builds on existing PLM capabilities.
The paper tackles the problem of obtaining high-quality sentence embeddings from pretrained language models without labeled data by generating datasets from scratch using large PLMs and finetuning smaller models, achieving outperformance on semantic textual similarity datasets.
To obtain high-quality sentence embeddings from pretrained language models (PLMs), they must either be augmented with additional pretraining objectives or finetuned on a large set of labeled text pairs. While the latter approach typically outperforms the former, it requires great human effort to generate suitable datasets of sufficient size. In this paper, we show how PLMs can be leveraged to obtain high-quality sentence embeddings without the need for labeled data, finetuning or modifications to the pretraining objective: We utilize the generative abilities of large and high-performing PLMs to generate entire datasets of labeled text pairs from scratch, which we then use for finetuning much smaller and more efficient models. Our fully unsupervised approach outperforms strong baselines on several semantic textual similarity datasets.