English(EN) Closed-Form Noise Calibration Against Membership Inference for Random-Allocation DP-SGD

新的 DP-SGD 噪声校准为对抗成员推断攻击提供了更紧密的界限

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-07 08:23

研究人员开发了一种新的、高效的方法来校准 DP-SGD（差分隐私随机梯度下降）训练中的噪声。该方法提供了一个单行公式来界定成员推断攻击（MIA）在训练模型上的准确性，与现有方法相比，显著减少了所需的噪声。该公式源于高斯分布和高斯混合分布之间的卡方散度，可以快速计算最优噪声乘数，从而提高测试准确性并缩短校准时间。 AI

影响在不牺牲准确性的前提下提高了机器学习模型的隐私保证，可能促使对敏感数据训练的更广泛采用。

排序理由学术论文，详细介绍了一种新的隐私保护机器学习方法。[lever_c_demoted from research: ic=1 ai=1.0]

在 Hugging Face Daily Papers 阅读 →

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: 0 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 学术论文，详细介绍了一种新的隐私保护机器学习方法。[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: 5 days old; Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.

完整方法见我们的编辑标准。

报道来源 [1]

Hugging Face Daily Papers TIER_1 English(EN) · 2026-10-07 08:23

针对随机分配差分隐私随机梯度下降的闭式差分隐私噪声校准

DP-SGD protects training data by adding Gaussian noise to clipped gradients. The amount of noise is usually chosen by running a numerical privacy accountant inside a search. We study DP-SGD with random allocation, where each epoch uses every record once, at a randomly chosen step…

本文如何被排名

报道来源 [1]

针对随机分配差分隐私随机梯度下降的闭式差分隐私噪声校准

相关实体

相关话题