Fetching the paper…

Learning Compositional Representation for Few-shot Visual Question Answering · Around