BERT is Not an Interlingua and the Bias of Tokenization

BERT is Not an Interlingua and the Bias of Tokenization Jasdeep Singh author Bryan McCann author Richard Socher author Caiming Xiong author 2019-11 text Proceedings of the 2nd Workshop on Deep Learning Approaches for Low-Resource NLP (DeepLo 2019) Colin Cherry editor Greg Durrett editor George Foster editor Reza Haffari editor Shahram Khadivi editor Nanyun Peng editor Xiang Ren editor Swabha Swayamdipta editor Association for Computational Linguistics Hong Kong, China conference publication singh-etal-2019-bert 10.18653/v1/D19-6106 https://aclanthology.org/D19-6106/ 2019-11 47 55