Robust Metric Learning based on the Rescaled Hinge Loss

Re-assign community