提出新方法，在压缩模型时同时保护抗攻击能力。
Adaptive Pruning with Module Robustness Sensitivity: Balancing Compression and Robustness
- 基于模块对扰动的敏感度动态决定剪枝策略。
- 在多个数据集上剪枝后仍保持高鲁棒性和准确率。
- 适合需要兼顾压缩与安全性的实际部署场景。
神经网络剪枝传统上依赖权重准则实现模型压缩，常忽视对抗鲁棒性与准确率之间的平衡。现有方法往往无法在剪枝后保持鲁棒性，使模型更易受对抗攻击。本文提出模块鲁棒性敏感度（MRS），一种量化层级对对抗扰动敏感度的新指标，并据此动态指导剪枝决策。基于MRS，我们提出模块鲁棒剪枝与微调（MRPF）算法，兼容任意对抗训练方法，兼具灵活性与可扩展性。在SVHN、CIFAR和Tiny-ImageNet上的大量实验表明，MRPF在多种架构（如ResNet、VGG、MobileViT）下显著提升对抗鲁棒性，同时保持优异的准确率与计算效率。此外，MRPF在鲁棒性、准确率与压缩率的平衡上持续优于当前最优结构化剪枝方法。本工作建立了一个实用且通用的鲁棒剪枝框架，解决了模型压缩与鲁棒性保留之间的长期权衡问题。
原文摘要 · Abstract (English)
Neural network pruning has traditionally focused on weight-based criteria to achieve model compression, frequently overlooking the crucial balance between adversarial robustness and accuracy. Existing approaches often fail to preserve robustness in pruned networks, leaving them more susceptible to adversarial attacks. This paper introduces Module Robustness Sensitivity (MRS), a novel metric that quantifies layer-wise sensitivity to adversarial perturbations and dynamically informs pruning decisions. Leveraging MRS, we propose Module Robust Pruning and Fine-Tuning (MRPF), an adaptive pruning algorithm compatible with any adversarial training method, offering both flexibility and scalability. Extensive experiments on SVHN, CIFAR, and Tiny-ImageNet across diverse architectures, including ResNet, VGG, and MobileViT, demonstrate that MRPF significantly enhances adversarial robustness while maintaining competitive accuracy and computational efficiency. Furthermore, MRPF consistently outperforms state-of-the-art structured pruning methods in balancing robustness, accuracy, and compression. This work establishes a practical and generalizable framework for robust pruning, addressing the long-standing trade-off between model compression and robustness preservation.
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。