Our paper on optimizing rank-based metrics with blackbox differentiation received flawless reviews and got accepted at CVPR 2020
Rank-based metrics are some of the most widely used criteria for performance evaluation of computervision models. Despite years of effort, direct optimization for these metrics remains a challengedue to their non-differentiable and non-decomposable nature. We present an efficient, theoreticallysound, and general method for differentiating rank-based metrics with mini-batch gradient descent.In addition, we address optimization instability and sparsity of the supervision signal that both arisefrom using rank-based metrics as optimization targets. Resulting losses based on recall and AveragePrecision are applied to image retrieval and object detection tasks. We obtain performance thatis competitive with state-of-the-art on standard image retrieval datasets and consistently improveperformance of near state-of-the-art object detectors.
Unser Ziel ist es, die Prinzipien von Wahrnehmen, Lernen und Handeln in autonomen Systemen zu verstehen, die mit komplexen Umgebungen interagieren. Das Verständnis wollen wir nutzen, um künstliche intelligente Systeme zu entwickeln.