arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2603.01971stat.MLcs.LG
收藏

LOCUS:一种分布无关的损失分位数分数用于风险感知预测

LOCUS: A Distribution-Free Loss-Quantile Score for Risk-Aware Predictions

Department of Statistics, Federal University of São Carlos(统计系,萨o 卡洛斯联邦大学);Institute of Mathematics and Computer Science, University of São Paulo(数学与计算机科学学院,圣保罗大学);School of Forest, Fisheries, and Geomatics Sciences, University of Florida(森林、渔业与地理科学学院,佛罗里达大学)

Matheus Barreto, Mário de Castro, Thiago R. Ramos, Denis Valle, Rafael Izbicki

AI总结:

LOCUS通过分布无关的损失分位数分数实现风险感知预测,有效降低大损失事件频率。

摘要:

现代机器学习模型在平均上可能准确,但仍然可能产生主导部署成本的错误。我们引入了Locus,一种分布无关的封装器,为固定预测函数生成每个输入的损失尺度可靠性分数。与量化标签的不确定性不同,Locus使用任何能够输出给定输入的损失预测分布的引擎来建模预测函数的实际损失。一个简单的拆分校准步骤将此函数转化为一个分布无关的可解释分数,该分数在不同输入之间具有可比性,并可解读为一个上限损失水平。该分数本身可用于排序,也可以通过阈值设置来获得透明的标记规则,从而实现对大损失事件的分布无关控制。在13个回归基准测试中,实验表明Locus在风险排序方面表现有效,并且比标准启发式方法减少了大损失频率。

英文摘要:

Modern machine learning models can be accurate on average yet still make mistakes that dominate deployment cost. We introduce Locus, a distribution-free wrapper that produces a per-input loss-scale reliability score for a fixed prediction function. Rather than quantifying uncertainty about the label, Locus models the realized loss of the prediction function using any engine that outputs a predictive distribution for the loss given an input. A simple split-calibration step turns this function into a distribution-free interpretable score that is comparable across inputs and can be read as an upper loss level. The score is useful on its own for ranking, and it can optionally be thresholded to obtain a transparent flagging rule with distribution-free control of large-loss events. Experiments across 13 regression benchmarks show that Locus yields effective risk ranking and reduces large-loss frequency compared to standard heuristics.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑