提出无需训练的去偏方法，让小模型也能高效精准排序。
Learning from Emptiness: De-biasing Listwise Rerankers with Content-Agnostic Probability Calibration
- 用无内容占位符估计位置偏置，通过自适应对比机制修正得分。
- 在10个基准上使0.6B小模型的NDCG提升超10点，超越聚合与数据增强方法。
- 无需额外训练或推理开销，适合资源受限场景下的高效排序应用。
生成式列表重排序利用全局上下文实现更优检索，但存在固有位置偏置问题，即模型对输入顺序敏感而与相关性无关。现有缓解方法面临两难：推理时聚合导致延迟过高，训练方法常无法消除深层先验，尤其在轻量级模型中表现不佳。为此，我们提出CapCal（内容无关概率校准）——一种无需训练的框架，机械解耦位置偏置与排序决策。通过内容无关占位符估计偏置分布，再以熵自适应对比机制修正输出logits。在10个基准上的评估表明，CapCal在无需训练的方法中表现最优，同时保持单次通过效率。尤为显著的是，它释放了轻量模型（如0.6B）的潜在能力，绝对NDCG提升超过10点，优于基于排列的聚合和数据增强基线。
原文摘要 · Abstract (English)
Generative listwise reranking leverages global context for superior retrieval but is plagued by intrinsic position bias, where models exhibit structural sensitivity to input order independent of relevance. Existing mitigations present a dilemma: inference-time aggregation incurs prohibitive latency, while training-based methods often fail to eradicate ingrained priors, particularly in compact models. To resolve this dilemma, we propose CapCal (Content-Agnostic Probability Calibration), a training-free framework that mechanically decouples positional bias from ranking decisions. By estimating the bias distribution via content-free placeholders, CapCal rectifies output logits through an entropy-adaptive contrastive mechanism. Evaluations across 10 benchmarks confirm that CapCal achieves superior performance among training-free methods while preserving single-pass efficiency. Notably, it unlocks the latent potential of lightweight models (e.g., 0.6B), delivering absolute NDCG gains exceeding 10 points and outperforming both permutation-based aggregation and data-augmentation 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。