Simultaneous Similarity-based Self-Distillation for Deep Metric Learning

     Karsten Roth 1 2 Timo Milbich 2 Bjorn Ommer 2 Joseph Paul Cohen * 3 Marzyeh Ghassemi * 4 1

             Abstract                  that embedding space dimensionality can be a driver for
                                   generalization across objectives due to higher representa-
   Deep Metric Learning (DML) provides a cru-            tion capacity. Indeed, this insight can be linked to recent
   cial tool for visua ...