1 paper
Lydia Nishimwe, Benoît Sagot, Rachel Bawden
NLP models have been known to perform poorly on user-generated content (UGC), mainly because it presents a lot of lexical variations and deviates from the standard texts on which m…