Leveraging Group Classification with Descending Soft Labeling for Deep Imbalanced Regression
Ruizhi Pu, Gezheng Xu, Ruiyi Fang, Binkun Bao, Charles X. Ling, Boyu Wang
TL;DR
This work tackles deep imbalanced regression (DIR) by reframing the problem as a divide-and-conquer task that couples group-level classification with group-specific regression. It introduces an ordinal group-aware contrastive loss to learn continuity-friendly representations, a multi-experts regression framework to handle different groups, and a symmetric descending soft-labeling scheme to preserve intrinsic label similarities during group classification. The final objective combines these components into L_final = L_grc + λ1 L_mse + λ2 L_soft, yielding state-of-the-art results on multiple DIR benchmarks and robust performance across majority, median, and minority groups. The approach demonstrates that explicitly leveraging classification to support regression via group structure and soft-label guidance can significantly mitigate imbalanced effects in regression tasks with continuous targets.
Abstract
Deep imbalanced regression (DIR), where the target values have a highly skewed distribution and are also continuous, is an intriguing yet under-explored problem in machine learning. While recent works have already shown that incorporating various classification-based regularizers can produce enhanced outcomes, the role of classification remains elusive in DIR. Moreover, such regularizers (e.g., contrastive penalties) merely focus on learning discriminative features of data, which inevitably results in ignorance of either continuity or similarity across the data. To address these issues, we first bridge the connection between the objectives of DIR and classification from a Bayesian perspective. Consequently, this motivates us to decompose the objective of DIR into a combination of classification and regression tasks, which naturally guides us toward a divide-and-conquer manner to solve the DIR problem. Specifically, by aggregating the data at nearby labels into the same groups, we introduce an ordinal group-aware contrastive learning loss along with a multi-experts regressor to tackle the different groups of data thereby maintaining the data continuity. Meanwhile, considering the similarity between the groups, we also propose a symmetric descending soft labeling strategy to exploit the intrinsic similarity across the data, which allows classification to facilitate regression more effectively. Extensive experiments on real-world datasets also validate the effectiveness of our method.