We present an overview of techniques that mitigate deep learning data hunger and (attempt to) leverage the power of deep learning on smaller datasets. Specifically, we will review techniques such as zero-shot learning, transfer-learning and the Transformer architecture.
