研究人员为训练 Kolmogorov-Arnold 网络（KANs，一种结构化的 MLP 替代方案）建立了新的理论界限。该工作分析了使用小批量随机梯度下降（SGD）训练的 KANs，包括具有相关噪声的差分隐私变体。这些发现揭示了非私有和私有训练模式之间的差距，表明差分隐私需要对数多项式网络宽度。 AI
影响 为 KANs 建立了理论基础，可能指导未来在隐私保护机器学习领域的研究。
排序理由 该集群包含两篇学术论文，详细介绍了特定类型神经网络架构（KANs）的理论分析和界限及其训练动态，包括隐私考虑因素。
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
arXiv:2605.12648v1 Announce Type: cross Abstract: We establish the first population risk bounds for Kolmogorov-Arnold Networks (KANs) trained by mini-batch SGD with gradient clipping, covering non-private SGD as well as differentially private SGD (DP-SGD) with Gaussian perturbati…
arXiv:2601.22409v3 Announce Type: replace-cross Abstract: Kolmogorov--Arnold Networks (KANs) have recently emerged as a structured alternative to standard MLPs, yet a principled theory for their training dynamics, generalization, and privacy properties remains limited. In this pa…
We establish the first population risk bounds for Kolmogorov-Arnold Networks (KANs) trained by mini-batch SGD with gradient clipping, covering non-private SGD as well as differentially private SGD (DP-SGD) with Gaussian perturbations that interpolate between independent and tempo…