Statistical few-shot learning for large-scale classification via parameter pooling

Andrew Simpson, Semhar Michael

In large-scale few-shot learning for classification problems, often there are a large number of classes and few high-dimensional observations per class. Previous model-based method…