不靠对抗训练也能变强：用振荡学习让模型自动抗攻击
Neither Adversarial Training Nor Purification: Emergent Adversarial Robustness from Oscillatory Predictive Learning
- 用振荡神经元+自监督预训练，从架构本身提升鲁棒性
- 在CIFAR-10上达76.63%抗干扰准确率，接近主流方法
- 无需生成对抗样本，适合追求高效鲁棒性的研究者
计算机视觉中的对抗鲁棒性通常依赖对抗训练或测试时净化，二者均需大量计算。本文探索是否可通过网络结构与表示学习的先验偏置实现鲁棒性。提出振荡预测学习（OPL）框架，结合人工柯朗托振荡神经元（AKOrN）与基于X-PhiNet的自监督预训练。因默认检查点使用随机振荡器初始状态，故与多种可复现的随机防御方法对比，评估其在AutoAttack-rand下的表现。在CIFAR-10和CIFAR-100上，OPL在ℓ∞, ε=8/255，EoT K=20条件下分别取得76.63±0.76%和50.44%的鲁棒准确率，在CIFAR-10-C上也通过了抗扰动测试，结果具有竞争力。
原文摘要 · Abstract (English)
Adversarial robustness in computer vision is still largely achieved through adversarial training or test-time adversarial purification, both of which introduce significant computational overhead by generating adversarial examples during training or performing iterative denoising at test time. We study whether empirical robustness can instead emerge from architectural and representation-learning inductive biases. We introduce Oscillatory Predictive Learning (OPL), a two-stage framework that combines Artificial Kuramoto Oscillatory Neurons (AKOrN) with predictive self-supervised pretraining using X-PhiNet. Because our default checkpoint uses randomized initial oscillator states, we compare it with other randomized adversarial defense methods that provide precise, reproducible, and strong attack protocols. Experiments on CIFAR-10 and CIFAR-100, with additional corruption evaluation on CIFAR-10-C, demonstrate that our method achieves competitive results under the AutoAttack-rand evaluation protocol. On CIFAR-10 and CIFAR-100, OPL attains 76.63$\pm$0.76$\%$ and 50.44$\%$ robust accuracy, respectively, under $\ell_\infty$, $\epsilon=8/255$, AutoAttack-rand with EoT $K=20$.
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。