@inproceedings{costeffectiveselectionofpretrainingdata, title = {Cost-effective Selection of Pretraining Data: A Case Study of Pretraining BERT on Social Media}, author = {Xiang Dai and Sarvnaz Karimi and Ben Hachey and Cecile Paris}, year = {2020}, booktitle = {Findings of the Association for Computational Linguistics 2020}, eprint = {2010.01150}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2010.01150v1}, }