PromptBERT: Improving BERT Sentence Embeddings with Prompts
This work addresses the issue of suboptimal sentence embeddings in BERT for natural language processing applications, representing an incremental advancement over existing methods like SimCSE.
The authors tackled the problem of improving BERT sentence embeddings by proposing PromptBERT, a novel contrastive learning method that uses prompts and template denoising, achieving improvements of 2.29 and 2.58 points over SimCSE with BERT and RoBERTa in unsupervised settings.
We propose PromptBERT, a novel contrastive learning method for learning better sentence representation. We firstly analyze the drawback of current sentence embedding from original BERT and find that it is mainly due to the static token embedding bias and ineffective BERT layers. Then we propose the first prompt-based sentence embeddings method and discuss two prompt representing methods and three prompt searching methods to make BERT achieve better sentence embeddings. Moreover, we propose a novel unsupervised training objective by the technology of template denoising, which substantially shortens the performance gap between the supervised and unsupervised settings. Extensive experiments show the effectiveness of our method. Compared to SimCSE, PromptBert achieves 2.29 and 2.58 points of improvement based on BERT and RoBERTa in the unsupervised setting.