We present VQA-MHUG - a novel 49-participant dataset of multimodal human...
It is now established that modern neural language models can be successf...
We introduce temporally and contextually-aware models for the novel task...
We present work in progress on the temporal progression of compositional...
share