Generating proper embedding of sentences through an unsupervised way is beneficial to semantic matching and retrieval problems in real-world scenarios. This paper presents Representation ALchemy (RepAL), an extremely simple post-processing method that enhances sentence representations. The basic idea in RepAL is to de-emphasize redundant information of sentence embedding generated by pre-trained models. Through comprehensive experiments, we show that RepAL is free of training and is a plug-and-play method that can be combined with most existing unsupervised sentence learning models. We also conducted in-depth analysis to understand RepAL.
@article{arxiv.2305.07824,
title = {A Simple and Plug-and-play Method for Unsupervised Sentence Representation Enhancement},
author = {Lingfeng Shen and Haiyun Jiang and Lemao Liu and Shuming Shi},
journal= {arXiv preprint arXiv:2305.07824},
year = {2023}
}