问题与现有方法的局限
Large language models with retrieval-augmented generation systematically underutilize evidence in the middle of multi-document contexts, a phenomenon known as Lost in the Middle. Existing mitigation strategies, such as document reordering and attention steering, can reduce boundary bias but remain brittle in noisy settings, where inaccurate relevance estimation and spurious documents still obscure useful evidence in central positions.
我们提出的框架：DuQUAD
We propose DuQUAD, a multi-document QA framework that mitigates positional bias through dual-view reasoning and quality-aware evidence pruning.
核心设计：双视角推理
DuQUAD combines a Local Agent guided by a Structural Fusion Score with a Global Agent that reasons over the full document set, providing both targeted evidence recovery and broad contextual understanding.
质量感知的证据剪枝
The framework further filters candidate evidence using explicit quality scoring and refines it at the sentence level to reduce noise.
实验结果与优势
DuQUAD consistently outperforms strong Lost-in-the-Middle mitigation baselines, including:
- hybrid retrieval and reranking
- multi-agent reasoning
- context optimization methods
It improves accuracy by 3.0% on average over the strongest competing baseline and gold evidence recall by 9.4%, while substantially reducing computational cost compared to existing multi-agent frameworks.