Deep Multi-Modal Sets

Austin Reiter, Menglin Jia, Pu Yang +1

Many vision-related tasks benefit from reasoning over multiple modalities to leverage complementary views of data in an attempt to learn robust embedding spaces. Most deep learning…