RADAR动态防御检索污染，用图模型选可靠信息
RADAR: Defending RAG Dynamically against Retrieval Corruption
- 将可靠内容选择建模为图能量最小化问题，精确求解
- 在动态数据集上实现更强鲁棒性，存储开销极低
- 适合需要实时更新知识的搜索系统使用
尽管RAG系统在动态网络搜索中应用日益广泛，但时间波动加剧了其对对抗攻击的脆弱性。现有静态防御方法难以应对演变威胁，且在动态场景中存储成本过高。我们提出RADAR框架，将可靠上下文选择建模为基于图的能量最小化问题，通过最大流最小割算法精确求解。引入贝叶斯记忆节点，递归更新信念状态，而非存储原始历史文档，有效平衡抗攻击稳定性与真实知识演化的适应性。在新构建的动态数据集上的实验表明，相比基线方法，RADAR在保持极低存储开销的同时，显著提升了鲁棒性和响应质量。
原文摘要 · Abstract (English)
While RAG systems are increasingly deployed in dynamic web search, temporal volatility amplifies their vulnerability to adversarial attacks. Existing static-oriented defenses struggle to handle evolving threats and incur prohibitive storage costs in dynamic settings. We propose RADAR, a framework that models reliable context selection as a graph-based energy minimization problem, solved exactly via Max-Flow Min-Cut. By incorporating a Bayesian memory node, RADAR recursively updates a belief state instead of archiving raw historical documents, effectively balancing stability against attacks with adaptability to genuine knowledge shifts. Experiments on a novel dynamic dataset show that RADAR achieves superior robustness and response quality with minimal storage overhead compared to the baselines.
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。