SCAD: Super-Class-Aware Debiasing for Long-Tailed Semi-Supervised Learning

ICLR 2026 Conference Submission10525 Authors

Keywords: semi-supervised learning

Abstract:

In long-tailed semi-supervised learning (LTSSL), pseudolabeling often creates a vicious cycle of bias amplification,
a problem that recent state-of-the-art methods attempt to mitigate using logit adjustment (LA). However, their adjustment
schemes, inherited from LA, remain inherently hierarchyagnostic, failing to account for the semantic relationships between classes. In this regard, we identify a critical yet overlooked problem of intra-super-class imbalance, where a toxic
combination of high semantic similarity and severe local
imbalance within each super-class hinders effective LTSSL.
This problem causes the model to reinforce on its errors,
leading to representation overshadowing. To break this cycle, we propose Super-Class-Aware Debiasing (SCAD), a
new framework that performs a dynamic, super-class-aware
logit adjustment. SCAD leverages the latent semantic structure between classes to focus its corrective power on the most
confusable groups, effectively resolving the local imbalances.
Our extensive experiments validate that SCAD achieves new
state-of-the-art performance, demonstrating the necessity of
a super-class-aware approach for robust debiasing.

Primary Area: unsupervised, self-supervised, semi-supervised, and supervised representation learning

Submission Number: 10525