Stochastic smoothing of the top-K calibrated hinge loss for deep imbalanced classification - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2022

Stochastic smoothing of the top-K calibrated hinge loss for deep imbalanced classification

Abstract

In modern classification tasks, the number of labels is getting larger and larger, as is the size of the datasets encountered in practice. As the number of classes increases, class ambiguity and class imbalance become more and more problematic to achieve high top-1 accuracy. Meanwhile, Top-K metrics (metrics allowing K guesses) have become popular, especially for performance reporting. Yet, proposing top-K losses tailored for deep learning remains a challenge, both theoretically and practically. In this paper we introduce a stochastic top-K hinge loss inspired by recent developments on top-K calibrated losses. Our proposal is based on the smoothing of the top-K operator building on the flexible "perturbed optimizer" framework. We show that our loss function performs very well in the case of balanced datasets, while benefiting from a significantly lower computational time than the state-of-the-art top-K loss function. In addition, we propose a simple variant of our loss for the imbalanced case. Experiments on a heavy-tailed dataset show that our loss function significantly outperforms other baseline loss functions.
Fichier principal
Vignette du fichier
garcin22a.pdf (2.93 Mo) Télécharger le fichier
17930 (1).pdf (2.12 Mo) Télécharger le fichier
17930.pdf (2.12 Mo) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-03828747 , version 1 (25-10-2022)

Licence

Identifiers

Cite

Camille Garcin, Maximilien Servajean, Alexis Joly, Joseph Salmon. Stochastic smoothing of the top-K calibrated hinge loss for deep imbalanced classification. ICML 2022 - 39th International Conference on Machine Learning, Jul 2022, Baltimore, United States. pp.7208-7222. ⟨hal-03828747⟩
149 View
34 Download

Altmetric

Share

Gmail Mastodon Facebook X LinkedIn More