用两种剪枝方式构建轻量防御模型，兼顾速度与抗攻击能力
Two is Better than One: Efficient Ensemble Defense for Robust and Compact Models
- 基于不同重要性评分对同一模型做差异化剪枝，生成多样性子模型
- 在CIFAR-10和SVHN上达到顶尖抗扰性能，推理速度提升最高1.86倍
- 适合部署在手机等资源受限设备上的高鲁棒性模型应用
基于深度学习的计算机视觉系统通常采用复杂庞大的架构以提升性能，但在资源受限的移动端和边缘设备上部署时面临挑战。为此，剪枝、量化和矩阵分解等模型压缩技术被提出；然而，这些压缩后的模型往往极易受到对抗攻击。本文提出高效集成防御（EED）技术，通过基于不同剪枝重要性评分对单一基础模型进行多样化压缩，并增强集成多样性，实现高对抗鲁棒性与资源效率的统一。EED在推理阶段动态确定所需子模型数量，减少冗余计算，同时保持高鲁棒性。在CIFAR-10和SVHN数据集上，EED的表现优于现有对抗剪枝方法，推理速度最高提升1.86倍。结果表明，EED是资源受限环境下强大的防御方案。
原文摘要 · Abstract (English)
Deep learning-based computer vision systems adopt complex and large architectures to improve performance, yet they face challenges in deployment on resource-constrained mobile and edge devices. To address this issue, model compression techniques such as pruning, quantization, and matrix factorization have been proposed; however, these compressed models are often highly vulnerable to adversarial attacks. We introduce the \textbf{Efficient Ensemble Defense (EED)} technique, which diversifies the compression of a single base model based on different pruning importance scores and enhances ensemble diversity to achieve high adversarial robustness and resource efficiency. EED dynamically determines the number of necessary sub-models during the inference stage, minimizing unnecessary computations while maintaining high robustness. On the CIFAR-10 and SVHN datasets, EED demonstrated state-of-the-art robustness performance compared to existing adversarial pruning techniques, along with an inference speed improvement of up to 1.86 times. This proves that EED is a powerful defense solution in resource-constrained environments.
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。