arXiv cs.AI· Navodita Sharma, Abhishek Shetty, Max Simchowitz, Pulkit Agrawal·· 13 小时前AI 评分31
条件流模型将认知不确定性转化为表观偶然方差
Conditional Flow Models Convert Epistemic Uncertainty into Apparent Aleatoric Variance
AI 导读
条件流模型（CFM）生成的样本方差常大于真实条件分布，即使在近乎确定性的回归任务中也是如此。这种额外方差源于模型尚未解析对条件变量的依赖，实质是把拟合真实条件分布时的认知不确定性转化为生成样本中的表观偶然方差。该方差随模型对复杂目标的欠拟合而增大，并随额外训练或数据增加而减小，据此可设计覆盖真实回归目标的高覆盖率预测集。
正文 · 原文
Abstract:Conditional Flow Models (CFMs) are increasingly being used in place of deterministic predictors across machine learning, including for tasks that were classically formulated as regression. Given their success in such tasks, we ask the question: what useful information do the conditional distributions produced by CFMs carry beyond what a point predictor would provide?
We identify a systematic phenomenon: in many natural regimes, CFM samples routinely have larger variance than the true conditional distribution, even in nearly deterministic regression problems. This effect is most pronounced where the model has not resolved the dependence on the conditioning variable, and thus the additional variance can be viewed as converting epistemic uncertainty in fitting the ground truth conditional distribution into apparent aleatoric variance in the generated samples. Through controlled synthetic and real-data experiments, and comparisons with explicit smoothing estimators, we show that the induced variance tracks model underfitting of complex or hard-to-fit targets, and decreases with additional training or data. We provide a theoretical account of the phenomenon in a simplified setting and demonstrate a consequent practical use-case of conditional flow-matching: designing high-coverage predictive sets that cover the ground-truth regression target.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.29172 [cs.LG] |
| (or arXiv:2610.29172v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.29172 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Navodita Sharma [view email]
[v1]
Wed, 21 Oct 2026 04:02:35 UTC (5,422 KB)
来源：arXiv cs.AI · arxiv.org