Word embeddings trained on visual co-occurrences from Visual Genome and ImageNet complement GloVe and improve several vision-language tasks, while random embeddings match learned ones on those supervised tasks.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
ViCo: Word Embeddings from Visual Co-occurrences
Word embeddings trained on visual co-occurrences from Visual Genome and ImageNet complement GloVe and improve several vision-language tasks, while random embeddings match learned ones on those supervised tasks.