Proxy-based Deep Metric Learning (DML) learns deep representations by embedding images close to their class representatives (proxies), commonly with respect to the angle between them. However, this disregards the embedding norm, which can carry additional beneficial context such as class- or image-intrinsic uncertainty. In addition, proxy-based DML struggles to learn class-internal structures. To address both issues at once, we introduce non-isotropic probabilistic proxy-based DML. We model images as directional von Mises-Fisher (vMF) distributions on the hypersphere that can reflect image-intrinsic uncertainties. Further, we derive non-isotropic von Mises-Fisher (nivMF) distributions for class proxies to better represent complex class-specific variances. To measure the proxy-to-image distance between these models, we develop and investigate multiple distribution-to-point and distribution-to-distribution metrics. Each framework choice is motivated by a set of ablational studies, which showcase beneficial properties of our probabilistic approach to proxy-based DML, such as uncertainty-awareness, better-behaved gradients during training, and overall improved generalization performance. The latter is especially reflected in the competitive performance on the standard DML benchmarks, where our approach compares favorably, suggesting that existing proxy-based DML can significantly benefit from a more probabilistic treatment. Code is available at github.com/ExplainableML/Probabilistic_Deep_Metric_Learning.
翻译:以代理为基础的深米力学习(DML)通过将图像嵌入与其班级代表(代理人)关系密切的图像(代理人),通常在他们之间的角度上,来学习深刻的表达方式。然而,这忽略了嵌入规范,因为嵌入规范可以带来更多有益的环境,如阶级或图像内在的不确定性。此外,基于代理的DML为学习阶级内部结构而奋斗。为了同时解决这两个问题,我们引入了非社会不稳定的代理服务器(DML)。我们将图像建为可反映形象内在不确定性的超视比值(vMF)分布,从而反映真实的图像处理方式(vMF)的分布。此外,我们为阶级提供非宗教化的vMises-Fisher(nmFMF)分布,以更好地代表阶级内部结构结构结构结构的复杂差异。为了测量这些模型之间的代比比值距离,我们开发并调查多种基于分布至点和分销至分配方式的衡量尺度。每个框架的选择都受到一系列通货膨胀研究的驱动,这些研究展示了我们具有更准确性、更准确性、更精确性、更精确性、更精确性、更能、更能、更能、更能、更能、更能更能更能反映整个的DMDMDMDMDR-L在总体的、更能、更能、更能、更能、更能、更能、更能、更能、更能、更能、更能、更能、更能、更能、更能、更能、更能、更能、更能、更能、更能、更能、更能、更能、更能、更能、更能、更能、更能更能反映到更能、更能、更能、更能、更能、更能、更能、更能、更能、更能性、更能、更能、更能、更能、更能、更能、更能、更能能能能、更能、更能更能、更能、更能、更能、更能、更能、更能、更能、更能更能、更能、更能、更能更能、更能、更能、更能、更能更能、更能、更能、更能