研究人员推出了 PRBench，这是一个旨在标准化深度学习模型概率鲁棒性评估的新基准。该基准在准确性、鲁棒性、训练效率和泛化误差等多个指标上比较了各种对抗性训练（AT）和针对性概率鲁棒性（PR）的训练方法。研究结果表明，AT 方法在提高对抗性和概率鲁棒性方面更为通用，而 PR 目标方法则提供了更好的泛化能力和干净准确性。此外，一个使用离散连续性模（DMOC）的新框架提供了一种数据驱动的方法来评估神经网络的鲁棒性，超越了传统的 Lipschitz 连续性度量，并在 ImageNet 等大型数据集上被证明是有效的。 AI
影响 新的基准和数据驱动的框架正在涌现，以更好地评估和提高人工智能模型在各种扰动下的可靠性。
排序理由 该集群包含多篇学术论文，介绍了用于评估人工智能模型鲁棒性的新基准和方法论。
AI 生成摘要 · Google Gemini · 来自 4 个来源。 我们如何撰写摘要 →
arXiv:2511.01724v3 Announce Type: replace-cross Abstract: Deep learning models are notoriously vulnerable to imperceptible perturbations. Most existing research centers on adversarial robustness (AR), which evaluates models under worst-case scenarios by examining the existence of…
arXiv:2605.28729v1 Announce Type: new Abstract: Robustness of neural networks is commonly quantified via local or global Lipschitz constants. However, Lipschitz continuity can be overly coarse or overly restrictive as global robustness measure, failing to capture nuanced, data-de…
arXiv:2509.22446v2 Announce Type: replace-cross Abstract: Double robustness is a major selling point of semiparametric and missing data methodology. Its virtues lie in protection against partial nuisance misspecification and asymptotic semiparametric efficiency under correct nuis…
Robustness of neural networks is commonly quantified via local or global Lipschitz constants. However, Lipschitz continuity can be overly coarse or overly restrictive as global robustness measure, failing to capture nuanced, data-dependent behavior. We propose a data-driven, arch…