Abstract: Learning generalized representations from limited training samples is crucial for applying deep neural networks in low-resource scenarios. Recently, methods based on contrastive ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results