Regularization via Adaptive Pairwise Label Smoothing

TL;DR

This paper proposes Pairwise Label Smoothing (PLS), a novel adaptive regularization technique that improves model generalization by smoothing label distributions based on sample pairs, outperforming traditional label smoothing methods.

Contribution

The paper introduces PLS, which learns pairwise label smoothing distributions during training, preserving label relationships and reducing overconfidence more effectively than existing methods.

Findings

PLS achieves up to 30% relative error reduction.

Models trained with PLS produce lower softmax scores.

PLS automatically learns smoothing distributions during training.

Abstract

Label Smoothing (LS) is an effective regularizer to improve the generalization of state-of-the-art deep models. For each training sample the LS strategy smooths the one-hot encoded training signal by distributing its distribution mass over the non ground-truth classes, aiming to penalize the networks from generating overconfident output distributions. This paper introduces a novel label smoothing technique called Pairwise Label Smoothing (PLS). The PLS takes a pair of samples as input. Smoothing with a pair of ground-truth labels enables the PLS to preserve the relative distance between the two truth labels while further soften that between the truth labels and the other targets, resulting in models producing much less confident predictions than the LS strategy. Also, unlike current LS methods, which typically require to find a global smoothing distribution mass through cross-validation search, PLS automatically learns the distribution mass for each input pair during training. We empirically show that PLS significantly outperforms LS and the baseline models, achieving up to 30% of relative classification error reduction. We also visually show that when achieving such accuracy gains the PLS tends to produce very low winning softmax scores.