English(EN) Security Properties of Neural Networks as Decision Problems

神经网络安全属性按复杂度分类

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-02 04:00

一篇新论文将神经网络的八个安全属性形式化为决策问题，并对其计算复杂度进行了分类。研究表明，在ReLU网络上，非干扰和单调性等属性是co-NP-完全的，而后门检测是Sigma_2^P-完全的。研究还发现，对参数进行量化（如比特翻转攻击）会使即使是简单网络的验证也成为exists-R-完全问题。 AI

影响为理解和验证神经网络的安全性提供了理论框架，可能指导未来鲁棒AI的研究。

排序理由详细介绍神经网络理论属性和复杂度的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 13 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 详细介绍神经网络理论属性和复杂度的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, safety; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Breaking (< 6h); Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

arXiv cs.LG TIER_1 English(EN) · Adrian Wurm · 2026-10-02 04:00

神经网络作为决策问题的安全性属性

arXiv:2609.39768v1 Announce Type: cross Abstract: Certifying a deployed neural network raises decision problems that the verification literature has not classified: whether the model carries a backdoor planted in its training data, whether a fault in its stored parameters can dri…