A simple cross-modal adaptation approach that learns from few-shot examples spanning different modalities. We demonstrate that one can indeed build a better visual dog classifier by reading about dogs ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results