---
id: 20260314-T0-18
title: "神经算子PDE代理的结构化不确定性量化"
title_en: "Structure-Aware Uncertainty Quantification for Neural Operator PDEs"
url: https://ai.daily.yangsir.net/daily/20260314-T0-18
issue_date: 2026-03-14
publish_date: 2026-03-13T04:00:00.000Z
source_name: "arXiv cs.LG (ML)"
source_url: https://arxiv.org/abs/2603.11052
---

# 神经算子PDE代理的结构化不确定性量化

arXiv论文提出了一种结构化认知不确定性量化方法，针对神经算子在PDE代理预测中的不确定性问题。该方法考虑有限数据、优化不完美和分布偏移等因素，提高了预测可靠性。

## English Version

**Structure-Aware Uncertainty Quantification for Neural Operator PDEs**

The arXiv paper proposes a structure-aware epistemic uncertainty quantification method for neural operator PDE surrogates. It addresses uncertainty from finite data, imperfect optimization, and distribution shift, improving prediction reliability.

---

**来源**：[arXiv cs.LG (ML)](https://arxiv.org/abs/2603.11052)

**详情页**：https://ai.daily.yangsir.net/daily/20260314-T0-18

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*