Skip to content

Latest commit

 

History

History
4 lines (4 loc) · 252 Bytes

File metadata and controls

4 lines (4 loc) · 252 Bytes

Motivation: embedding triplet and multimodal data into vector space
Methods: encoders: multimodal data into vector; decoders: generate multi-modal values
Datasets: MovieLens-100k, YAGO-10
Experiments: link prediction, generating text and images