← 研踪 PaperTrailarXiv:2603.11750 · v3
Trustworthy Calibration or Vulnerability? Unveiling the Over-Calibration Paradox and Attack Surface in Rashomon Sets
首次提交 2026-03-12 17:54
版本更新 2026-11-16 23:12
雷达观测 2026-11-17 16:10
摘要
Post-hoc calibration is standard practice for aligning classifier confidence scores with empirical frequencies in high-stakes decision systems. However, its downstream effects on decision stability across competing, equally accurate models remain largely unexamined. In this paper, we formalize the over-calibration paradox within the framework of predictive multiplicity, reflecting the Rashomon effect where competing models with near-identical loss assign conflicting predictions to the same individual. Evaluating nine credit risk portfolios, we examine the impact of Platt Scaling, Isotonic Regression, and Beta Calibration on individual prediction agreement and expected calibration error. We identify two distinct operating regimes. In systematically miscalibrated models, post-hoc calibration consistently aligns probability estimates and reduces prediction disagreement. Conversely, when baseline models are already well-calibrated, post-hoc calibration, particularly non-parametric estimators constrained by finite sample sizes, overfits to hold-out noise. This distortion shifts decision thresholds near the classification boundary, paradoxically increasing both expected calibration error and predictive multiplicity. From an auditing and security standpoint, this paradox exposes a practical vulnerability to soft adversarial manipulation, whereby an interested party or automated model search can exploit calibration choices to steer specific individual outcomes while preserving aggregate performance and passing black-box evaluations.
作者备注:16 pages, 2 figures
观测到的版本
相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。
v3 · 2026-11-16 23:12 · 最新观测版本
Trustworthy Calibration or Vulnerability? Unveiling the Over-Calibration Paradox and Attack Surface in Rashomon Sets
Post-hoc calibration is standard practice for aligning classifier confidence scores with empirical frequencies in high-stakes decision systems. However, its downstream effects on decision stability across competing, equally accurate models remain largely unexamined. In this paper, we formalize the over-calibration paradox within the framework of predictive multiplicity, reflecting the Rashomon effect where competing models with near-identical loss assign conflicting predictions to the same individual. Evaluating nine credit risk portfolios, we examine the impact of Platt Scaling, Isotonic Regression, and Beta Calibration on individual prediction agreement and expected calibration error. We identify two distinct operating regimes. In systematically miscalibrated models, post-hoc calibration consistently aligns probability estimates and reduces prediction disagreement. Conversely, when baseline models are already well-calibrated, post-hoc calibration, particularly non-parametric estimators constrained by finite sample sizes, overfits to hold-out noise. This distortion shifts decision thresholds near the classification boundary, paradoxically increasing both expected calibration error and predictive multiplicity. From an auditing and security standpoint, this paradox exposes a practical vulnerability to soft adversarial manipulation, whereby an interested party or automated model search can exploit calibration choices to steer specific individual outcomes while preserving aggregate performance and passing black-box evaluations.
查看此版本解读 →查看此版本原文 ↗