---
id: 20260825-T0-05
title: "神经PDE算子遭“错误物理”后门攻击，数据投毒难被验证察觉"
title_en: "New 'Wrong-Physics' Backdoor Attack Targets Neural PDE Operators in Solver Archives"
url: https://ai.daily.yangsir.net/daily/20260825-T0-05
issue_date: 2026-08-25
publish_date: 2026-08-24T04:00:00.000Z
category: research
source_name: "arXiv cs.LG (ML)"
source_url: https://arxiv.org/abs/2608.20439
---

# 神经PDE算子遭“错误物理”后门攻击，数据投毒难被验证察觉

研究人员提出一种针对神经PDE算子的数据投毒攻击方法——跨参数重连。当模型利用可复用的求解器存档进行训练时，攻击者刻意构造“真实”但由错误物理规则生成的数据样本。由于模型验证常依赖于预测误差和与参数无关的合理性检查，这种带有错误物理的后门很难被察觉，可能导致模型解决特定问题时产生灾难性偏差。

## English Version

**New 'Wrong-Physics' Backdoor Attack Targets Neural PDE Operators in Solver Archives**

Researchers introduced 'cross-parameter relinking', a new data-poisoning backdoor attack targeting neural PDE operators trained on reusable solver archives. By injecting carefully crafted data points that adhere to 'wrong physics', the attack compromises the model's integrity. Standard validation methods, including prediction error and parameter-agnostic plausibility checks, often fail to detect these backdoors. As a result, models can produce catastrophically incorrect predictions for specific problems, posing a serious risk for scientific machine learning applications.

---

**来源**：[arXiv cs.LG (ML)](https://arxiv.org/abs/2608.20439)

**详情页**：https://ai.daily.yangsir.net/daily/20260825-T0-05

---

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