arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2203.08547cs.CV
收藏

基于代理的深度度量学习的非各向同性正则化

Non-isotropy Regularization for Proxy-based Deep Metric Learning

University of Tübingen(图宾根大学);DeepMind(深度思维);MPI for Intelligent Systems(马克斯·普朗克智能系统研究所)

Karsten Roth, Oriol Vinyals, Zeynep Akata

摘要:

深度度量学习(DML)旨在学习表示空间,在这些空间上,语义关系可以通过预定义的距离度量简单地表达。性能最好的方法通常利用类代理作为样本替身,以实现更好的收敛和泛化。然而,这些代理方法仅针对样本-代理距离进行优化。考虑到所使用的距离函数的固有非双射性,这可能会导致局部各向同性的样本分布,从而导致由于难以解析样本之间的局部结构和类内关系而错过关键的语义上下文。为了缓解这个问题,我们提出了用于基于代理的深度度量学习的非各向同性正则化($\mathbb{NIR}$)。通过利用规范化流,我们从各自的类代理中强制执行样本的独特可翻译性。这使我们能够明确地诱导样本在代理周围的非各向同性分布以进行优化。在这样做的过程中,我们配备了基于代理的目标,以更好地学习本地结构。广泛的实验突出了$\mathbb{NIR}$的一致泛化优势,同时在标准基准CUB 200 -2011,Cars 196和斯坦福在线产品上实现了具有竞争力的和最先进的性能。此外,我们发现基于代理的方法的优越的收敛性能仍然被保留,甚至改进,使$\mathbb{NIR}$非常有吸引力的实际使用。代码可在https://github.com/ExplainableML/NonIsotropicProxyDML获得。

英文摘要:

Deep Metric Learning (DML) aims to learn representation spaces on which semantic relations can simply be expressed through predefined distance metrics. Best performing approaches commonly leverage class proxies as sample stand-ins for better convergence and generalization. However, these proxy-methods solely optimize for sample-proxy distances. Given the inherent non-bijectiveness of used distance functions, this can induce locally isotropic sample distributions, leading to crucial semantic context being missed due to difficulties resolving local structures and intraclass relations between samples. To alleviate this problem, we propose non-isotropy regularization ($\mathbb{NIR}$) for proxy-based Deep Metric Learning. By leveraging Normalizing Flows, we enforce unique translatability of samples from their respective class proxies. This allows us to explicitly induce a non-isotropic distribution of samples around a proxy to optimize for. In doing so, we equip proxy-based objectives to better learn local structures. Extensive experiments highlight consistent generalization benefits of $\mathbb{NIR}$ while achieving competitive and state-of-the-art performance on the standard benchmarks CUB200-2011, Cars196 and Stanford Online Products. In addition, we find the superior convergence properties of proxy-based methods to still be retained or even improved, making $\mathbb{NIR}$ very attractive for practical usage. Code available at https://github.com/ExplainableML/NonIsotropicProxyDML.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑