Adaptive Margin Ordinal Loss: Penalizing Center-Class Hedging in Ordinal Classification

Abstract

Standard cross-entropy loss causes neural networks trained on ordinal classification tasks to hedge predictions toward center classes, a failure mode we term \emph{center-class hedging}. This occurs because predicting the middle class minimizes expected symmetric loss, making it the path of least resistance regardless of the true label. Existing ordinal losses address related problems such as large-error penalization and rank consistency, but none directly suppresses center-class hedging as a function of where the true label lies relative to the ordinal center. We propose the Adaptive Margin Ordinal Loss (AMOL), a multiplicative weight applied to per-class loss terms of the form $m(k,y) = 1 + \alpha \cdot (1 - |k-c|/c) \cdot (|y-c|/c)$, where $c$ is the center class, $k$ is the candidate class, and $y$ is the true label. The weight encodes a joint condition: it is large only when the candidate class is near center and the true label is far from center, collapsing to standard behavior otherwise. We further introduce the Center-Hedging Rate (CHR) as a diagnostic metric that directly quantifies this failure mode. Across four ordinal classification benchmarks and five random seeds, AMOL achieves the best or tied-best Quadratic Weighted Kappa (QWK) on all four datasets compared to cross-entropy, OLL, and SORD baselines. An asymmetric variant (AMOL-asym) eliminates center-class hedging entirely on the Abalone dataset ($\text{CHR} = 0.000 \pm 0.000$ across all five seeds, $n \approx 266$ extreme-class test samples per run), compared to $0.074 \pm 0.005$ for standard cross-entropy.

Authors 1

Manisha Kandel

Specification

Official page

arxiv.org/abs/2609.10752

Source:arXiv (Atom API + RSS)·T1·observed 41 min ago·high

Arxiv announce type

new

Source:arXiv (Atom API + RSS)·T1·observed 41 min ago·high

arXiv id

2609.10752

Source:arXiv (Atom API + RSS)·T1·observed 41 min ago·high

Categories

cs.LG, cs.AI

Source:arXiv (Atom API + RSS)·T1·observed 41 min ago·high

PDF

arxiv.org/pdf/2609.10752

Source:arXiv (Atom API + RSS)·T1·observed 41 min ago·high

Primary category

cs.LG

Source:arXiv (Atom API + RSS)·T1·observed 41 min ago·high

Published

11 Sept 2026

Source:arXiv (Atom API + RSS)·T1·observed 41 min ago·high

Each value shows its source, tier and observation time. Conflicting claims are kept side by side and flagged — never averaged. How AI Atlas records facts →

Provenance