研究人员开发了一个新的认证鲁棒回归框架，解决了现有方法的局限性。这种新颖的方法提供了一个以预测为中心的证书，确保了平滑模型预测的稳定性，并且便于实时计算。在MNIST旋转任务上的实验表明，与alpha-smoothing等当前最先进的技术相比，结合梯度信息可以获得更严格的鲁棒性证书。 AI
影响 这项研究可能带来更可靠、更稳定的回归模型，尤其是在对抗性环境中。
排序理由 该集群包含一篇详细介绍机器学习新方法的学术论文。
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
arXiv:2607.05536v1 Announce Type: new Abstract: Randomized smoothing has emerged as a scalable technique for certifying the adversarial robustness of classifiers. However, its application to regression remains under-explored and faces unique challenges. Existing regression certif…
Randomized smoothing has emerged as a scalable technique for certifying the adversarial robustness of classifiers. However, its application to regression remains under-explored and faces unique challenges. Existing regression certificates rely on probabilistic acceptance regions …