ALIGN-MLM: Word Embedding Alignment is Crucial for Multilingual Pre-training.

Henry Tang Ameet Deshpande Karthik Narasimhan

Published in: CoRR (2022)

Keyphrases

word level
word alignment
training set
language specific
language independent
co occurrence
training corpus
training process
training samples
digital libraries
training examples
n gram
cross language
dynamic time warping
sequence alignment
statistical machine translation
supervised learning
parallel corpus