Addressing Long-Tail Noisy Label Learning Problems: a Two-Stage Solution with Label Refurbishment Considering Label Rarity

arXiv:2403.02363v1 Announce Type: new
Abstract: Real-world datasets commonly exhibit noisy labels and class imbalance, such as long-tailed distributions. While previous research addresses this issue by differentiating noisy and clean samples, reliance on information from predictions based on noisy long-tailed data introduces potential errors. To overcome the limitations of prior works, we introduce an effective two-stage approach by combining soft-label refurbishing with multi-expert ensemble learning. In the first stage of robust soft label refurbishing, we acquire unbiased features through contrastive learning, …

abstract arxiv class cs.ai cs.lg data datasets errors information issue labels predictions reliance research samples solution stage type world