
Jung-Woo Ha
@jungwooha2
PhD, VP /
Research Fellow, #AI_LAB @ #NAVER Cloud/
Social Chair@ICML24 /
Member, #NAEK
ID: 1144194850746540032
https://scholar.google.com/citations?user=eGj3ay4AAAAJ&hl=ko 27-06-2019 10:44:45
728 Tweet
3,3K Takipçi
2,2K Takip Edilen

Happy to share our new #BLEURT-based reward optimization using a ranking-based contrastive margin loss for #NMT. We reconfirm the limitation of BLEU from our reward optimization results and human evaluation. Congrats to Raphael Shu, Kangmin Yoo. arxiv: arxiv.org/abs/2104.07541
