验证方法效果依赖数据集，训练策略需因数据而异。
Veriphi: Attack-Guided Neural Network Verification with Dataset-Dependent Training Methods
- 用攻击引导的验证方法加速形式化认证
- IBP在MNIST上达78%认证准确率，但对CIFAR-10无效
- 适用于航空航天等高安全场景的大型模型验证
我们提出Veriphi，一个基于GPU加速的神经网络验证系统，结合快速对抗攻击与alpha,beta-CROWN形式边界认证。在MNIST和CIFAR-10上，采用三种训练方法（标准、对抗、认证）进行系统实验，发现训练方法的有效性根本上依赖于数据集特性。区间传播（IBP）在简单数据集MNIST（784维）上实现78%的认证准确率，但在更复杂的CIFAR-10上几乎无法提供有效认证；而PGD对抗训练在小扰动下表现突出，达到94%的认证准确率。通过攻击引导的伪证方法实现5倍验证速度提升，并成功扩展至105.8M参数的生产级模型，应用于真实航空航天物流优化任务。结果挑战了‘认证训练始终优于对抗训练’的假设，表明验证策略选择需高度依赖具体上下文。
原文摘要 · Abstract (English)
We present Veriphi, a GPU-accelerated neural network verification system that combines fast adversarial attacks with formal bound certification using alpha,beta-CROWN methods. Through systematic experiments on MNIST and CIFAR-10 using three training methodologies (standard, adversarial, certified), we demonstrate that training method effectiveness is fundamentally dataset-dependent. Interval Bound Propagation (IBP) achieves 78% certified accuracy on simple MNIST (784 dimensions) but provides negligible certification performance on the more complex CIFAR-10 dataset, where PGD adversarial training dominates with 94% certification at small perturbations. We achieve 5x verification speedup through attack-guided falsification and scale our approach to production-size models (105.8M parameters) for real-world aerospace logistics optimization. Our results challenge the assumption that certified training universally outperforms adversarial training, showing context matters critically for verification strategy selection.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。