TLM: Token-Level Masking for Transformers.

Yangjun Wu, Kebin Fang, Dongxiang Zhang, Han Wang, Hao Zhang, Gang Chen

Browse the full EMNLP paper archive.