2 papers
stat.ML2021
ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision
Wonjae Kim, Bokyung Son, Ildoo Kim
Vision-and-Language Pre-training (VLP) has improved performance on various joint vision-and-language downstream tasks. Current approaches to VLP heavily rely on image feature extra…
cs.CL2020
Revisiting Modularized Multilingual NMT to Meet Industrial Demands
Sungwon Lyu, Bokyung Son, Kichang Yang +1
The complete sharing of parameters for multilingual translation (1-1) has been the mainstream approach in current research. However, degraded performance due to the capacity bottle…