The objective in extreme multi-label learning is to learn a classifier that can automatically tag a datapoint with the most relevant subset of labels from an extremely large label set. This competition  provides a relatively small dataset called EURLex-4K which has just ~4000 labels and ~15,000 training points.