paper-with-me

홈 › Papers

On Calibration in Multi-Distribution Learning

2024-12-18 · Rajeev Verma, Volker Fischer, Eric Nalisnick

Modern challenges of robustness, fairness, and decision-making in machine learning have led to the formulation of multi-distribution learning (MDL) frameworks in which a predictor is optimized across multiple distributions. We study the calibration properties of MDL to better understand how the predictor performs uniformly across the multiple distributions. Through classical results on decomposing proper scoring losses, we first derive the Bayes optimal rule for MDL, demonstrating that it maximizes the generalized entropy of the associated loss function. Our analysis reveals that while this approach ensures minimal worst-case loss, it can lead to non-uniform calibration errors across the multiple distributions and there is an inherent calibration-refinement trade-off, even at Bayes optimality. Our results highlight a critical limitation: despite the promise of MDL, one must use caution when designing predictors tailored to multiple distributions so as to minimize disparity.

📄 PDF Abstract BibTeX arXiv:2412.14142

Code (0)

등록된 구현이 없습니다.

Tasks

Decision MakingFairness

Methods 이 논문이 사용한 방법론

MDL Minimum Description Length provides a criterion for the selection of models, regardless of their complexity, without the restrictive assumption that the data form a sample…

Similar Papers 제목 키워드 기반

Calibrating Predictions to Decisions: A Novel Approach to Multi-Class Calibration

2021-07-12 · NeurIPS 2021 12 · Shengjia Zhao, Michael P. Kim, Roshni Sahoo, Tengyu Ma 외

When facing uncertainty, decision-makers want predictions they can trust. A machine learning provider can convey confidence to decision-makers by guaranteeing their predictions are distribution calibrated -- amongst the …

Decision Making

Bridging Multicalibration and Out-of-distribution Generalization Beyond Covariate Shift

2024-06-02 · Jiayun Wu, Jiashuo Liu, Peng Cui, Zhiwei Steven Wu

We establish a new model-agnostic optimization framework for out-of-distribution generalization via multicalibration, a criterion that ensures a predictor is calibrated across a family of overlapping groups. Multicalibra…

Out-of-Distribution Generalization

Score-based calibration testing for multivariate forecast distributions

2022-11-29 · Malte Knüppel, Fabian Krüger, Marc-Oliver Pohle

Calibration tests based on the probability integral transform (PIT) are routinely used to assess the quality of univariate distributional forecasts. However, PIT-based calibration tests for multivariate distributional fo…

Time SeriesTime Series Analysis

A Unifying Perspective on Multi-Calibration: Game Dynamics for Multi-Objective Learning

2023-02-21 · NeurIPS 2023 11

We provide a unifying framework for the design and analysis of multicalibrated predictors. By placing the multicalibration problem in the general setting of multi-objective learning -- where learning guarantees must hold…

Fairness

Parametric and Multivariate Uncertainty Calibration for Regression and Object Detection

2022-07-04 · Fabian Küppers, Jonas Schneider, Anselm Haselhoff

Reliable spatial uncertainty evaluation of object detection models is of special interest and has been subject of recent work. In this work, we review the existing definitions for uncertainty calibration of probabilistic…

object-detectionObject DetectionObject Trackingregression+1