给不可信AI代理装一道差分隐私闸门：用噪声淹没敏感数据，Claude实测演示如何在保护患者隐私时耗尽预算，开源可玩。
Article URL: https://github.com/yashmahajan10/llm-differential-privacy-gateway Comments URL: https://news.ycombinator.com/item?id=49119049 Points: 1 #…
PromptPET通过智能混淆保护提示词隐私，在不明显牺牲回答质量的前提下平衡安全与效用，适合处理敏感查询。
arXiv:2607.02932v1 Announce Type: cross Abstract: Privacy is an important challenge when users interact with AI chatbots, since users may share sensit…
差分隐私的粗粒度保护之外，这项研究提出用可预测性做更精细的隐私度量，值得关注。
arXiv:2606.20546v1 Announce Type: new Abstract: Differential privacy (DP) ensures rigorous individual-level privacy guarantees against even the most k…
一篇严谨的差分隐私高斯过程后验采样研究，包含主文8页+25页附录，理论深度与实用分析兼备
arXiv:2606.17995v1 Announce Type: cross Abstract: We study the privacy of releasing posterior sample paths from a Gaussian process (GP) when the entir…
一篇深度分析美国人口普查禁用差分隐私政策的博客，帮你快速理解“噪声注入”争议与隐私保护平衡难题。
Article URL: https://desfontain.es/blog/banning-noise.html Comments URL: https://news.ycombinator.com/item?id=48517377 Points: 505 # Comments: 261
无需牺牲隐私即可合成逼真数据，为敏感信息处理提供全新方案。
arXiv:2602.05833v2 Announce Type: replace Abstract: There is a need for synthetic training and test datasets that replicate statistical distributions …
重新审视低秩适应（LoRA）在私有LLM微调中的应用，探讨差分隐私与效率的平衡。
arXiv:2510.01137v3 Announce Type: replace Abstract: Privacy is a central concern when fine-tuning large language models (LLMs) on sensitive data, and …
面向大语言模型的高效差分隐私训练方法，提出随机裁剪机制优化DP-SGD，兼顾隐私与性能。
arXiv:2605.24879v1 Announce Type: new Abstract: Large language models (LLMs) are trained on vast datasets that may contain sensitive information. Diff…
新研究用可证明方式保护微调大模型免遭训练数据窃取，同时维持模型效能，隐私与实用兼得。
arXiv:2602.00688v2 Announce Type: replace Abstract: Fine-tuning large language models (LLMs) on sensitive datasets raises privacy concerns, as trainin…
一键直达差分隐私模型融合最新研究，arXiv免费提供全文下载与引用管理，助你紧跟学术前沿。
arXiv:2604.20985v2 Announce Type: replace Abstract: In machine learning, privacy requirements at inference or deployment time often evolve due to chan…
差分隐私+选择性微调，为大模型训练数据保护提供新方案。
arXiv:2605.17432v1 Announce Type: new Abstract: Large language models (LLMs) are commonly adapted to downstream tasks through fine-tuning, but fine-tu…
这篇论文评估大模型在差分隐私推理中的表现，为自动设计隐私保护算法提供新基准。
arXiv:2604.15851v2 Announce Type: replace-cross Abstract: Differential privacy (DP) has a wide range of applications for protecting data privacy, but …