基于体制感知的同行专业化方法，用于在异构知识冲突下实现稳健的检索增强生成
Regime-Aware Peer Specialization for Robust RAG under Heterogeneous Knowledge Conflicts
- School of Computer Science, Beijing Institute of Technology（北京理工大学计算机科学学院）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
提出RAPS-DA框架，通过样本级三体制划分（接地、仲裁、抵抗）和token级双选择器，解决RAG中异构知识冲突问题，在固定模型规模下超越多种基线。
AI中文摘要：
检索增强生成（RAG）通过将生成过程锚定在外部上下文中来改进语言模型。然而，当检索到的上下文与模型的参数化知识冲突时，它可能变得脆弱。这种冲突涵盖了一个可靠性谱系，从可靠和部分可靠的证据到对抗性上下文。现有的补救措施通常采用体制无关的监督来处理这种异构冲突，这可能会混淆不同可靠性体制下的不兼容学习信号。为了解耦这些信号，我们提出了RAPS-DA，一个体制感知的同行专业化框架，在两个互补的粒度上解决冲突。在样本级别，冲突被分为三个体制，包括接地、仲裁和抵抗，每个体制从共享的基础模型训练一个同尺度的同行专家。然后，每个样本被硬路由到其体制匹配的同行，进行策略上的反向KL监督。在token级别，一个双层选择器使用教师间分歧、学生-教师差异和学生熵来过滤无信息或不稳定的token，对自信不匹配的token进行上加权，并随着学生的成熟逐渐将监督集中在高冲突token上。收益来自于固定模型尺度上的专业化，而不是更强的教师，并且同行专家仅在训练期间存在，因此部署的学生不需要体制标签或同行访问。在五个冲突场景和两个分布外基准上的实验表明，RAPS-DA超越了所有提示、解码、微调、强化学习和单教师基线。
英文摘要：
Retrieval-augmented generation (RAG) improves language models by grounding generation in external context. However, it can be fragile when the retrieved context conflicts with the model's parametric knowledge. Such conflicts span a reliability spectrum, ranging from reliable and partially reliable evidence to adversarial context. Existing remedies often handle such heterogeneous conflicts with regime-agnostic supervision, which can conflate incompatible learning signals across reliability regimes. To disentangle these signals, we propose RAPS-DA, a regime-aware peer specialization framework that addresses conflict at two complementary granularities. At the sample level, conflicts are divided into three regimes, including Grounding, Arbitration, and Resistance, with one same-scale peer specialist trained per regime from a shared base model. Each sample is then hard-routed to its regime-matched peer for on-policy reverse-KL supervision. At the token level, a dual-layer selector uses inter-teacher disagreement, student-teacher divergence, and student entropy to filter uninformative or unstable tokens, upweight confidently misaligned ones, and gradually focus supervision on high-conflict tokens as the student matures. Gains stem from specialization at a fixed model scale, not from a stronger teacher, and the peer specialists exist only during training, so the deployed student requires no regime labels or peer access. Experiments on five conflict scenarios and two out-of-distribution benchmarks show RAPS-DA surpasses all prompting, decoding, fine-tuning, RL, and single-teacher baselines.
