Learning to learn by yourself: Unsupervised meta‐learning with self‐knowledge distillation for COVID‐19 diagnosis from pneumonia cases

To the best of our knowledge, our proposed data set is the largest data set compared to existing publicly available COVID‐19 data sets except for normal cases . It is worth noting that the proposed data set is an unbalanced data set, which is exactly consistent with the long‐tail problem solved by meta‐learning. In this section, we first describe how we built our proposed COVID‐19 Pneumonia Data set and introduce the structure of our proposed data set. Then, we can get the anchor images set X a , positive images set X p and negative images set X n by following the above strategy. On the COVID‐CT data set, we follow the data split of this data set.