Proxy Anchor Loss for Deep Metric Learning

By Sungyeon Kim, Dongwon Kim, Minsu Cho, Suha Kwak. Computer Vision and Pattern Recognition, 2020

468 citations recorded by the source corpus.

Existing metric learning losses can be categorized into two classes: pair-based and proxy-based losses. The former class can leverage fine-grained semantic relations between data points, but slows convergence in general due to its high training complexity. In contrast, the latter class enables fast and reliable convergence, but cannot consider the rich data-to-data relations. This paper presents a new proxy-based loss that takes advantages of both pair- and proxy-based methods and overcomes their limitations. Thanks to the use of proxies, our loss boosts the speed of convergence and is robust against noisy labels and outliers. At the same time, it allows embedding vectors of data to interact with each other in its gradients to exploit data-to-data relations. Our method is evaluated on four public benchmarks, where a standard network trained with our loss achieves state-of-the-art performance and most quickly converges.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/f7646cccbd6edbc148d08fea37e31bcd0592c992

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

Citations in the corpus

ETHOS: a multi-label hate speech detection dataset (2020, background): “Therefore, the adoption of metric learning mechanisms may help us alleviate the hubness phenomenon which put obstacles on recovering distinct classes [25].”
ContraGAN: Contrastive Learning for Conditional Image Generation (2020, background): “While the proxy-based losses [33, 38, 39] relieves mining complexity using trainable class embedding vectors, such losses do not explicitly take data-to-data relations [40] into account.”
Characterizing Generalization under Out-Of-Distribution Shifts in Deep Metric Learning (2021, methodology): “More recent work has tackled sampling complexity through the usage of proxy-representations utilized as sample stand-ins during training, following a NCA [16] objective [41, 26, 64], leveraging softmax-style training through class-proxies [8, 73] or simulating intraclass structures [46].”
ABO: Dataset and Benchmarks for Real-World 3D Object Understanding (2021, background): “Designs of losses for representation and metric learning in the recent literature [43] can be broadly summarized into: (i) pair-based losses [22, 44], where losses are computed based on 2, 3 or more distances between samples in the batches, with the goal to minimize the distances between the sample ”