论文标题
低损失子空间压缩，可用于针对多代理后门攻击的清洁收益
Low-Loss Subspace Compression for Clean Gains against Multi-Agent Backdoor Attacks
论文作者
论文摘要
对多代理后门攻击的最新探索表明了反击效应，这是对后门攻击的自然防御效果，后门攻击是随机分类的。这产生了低精度W.R.T.的副作用干净的标签，激发了本文在构建多代理后门防御的构建方面的工作，该防御能力使精度最大化W.R.T.清洁标签并最大程度地减少毒药标签的标签。我们建立在代理动力学和低损失子空间结构的基础上，我们贡献了三种防御能力，可提高多代理后门鲁棒性。
Recent exploration of the multi-agent backdoor attack demonstrated the backfiring effect, a natural defense against backdoor attacks where backdoored inputs are randomly classified. This yields a side-effect of low accuracy w.r.t. clean labels, which motivates this paper's work on the construction of multi-agent backdoor defenses that maximize accuracy w.r.t. clean labels and minimize that of poison labels. Founded upon agent dynamics and low-loss subspace construction, we contribute three defenses that yield improved multi-agent backdoor robustness.
