提出一种快速可解释的排序模型，仅在必要时生成理由。
OG-Rank: Learning to Rank Fast and Slow with Uncertainty and Reward-Trend Guided Adaptive Exploration
- 用单解码器结合置信度门控，先快速排序，仅当不确定时才生成解释。
- 在临床医嘱排序任务中，召回率最高达0.56，nDCG@20达0.699，门控触发率45%。
- 适合对实时性与可解释性都有要求的决策场景，部署简单且成本可控。
临床医生需要低延迟且能解释结果的排序系统。为满足基于解码器的快速重排序需求，我们提出OG-Rank，一种单解码器方法：通过合并首词评分信号与不确定性门控的解释步骤，在一次遍历中完成所有候选排序，并仅在列表真正模糊时生成简明结构化理由，确保延迟可预测。采用聚焦难例的课程训练策略，OG-Rank在会诊范围内的排序任务中表现优异（快速路径：Recall@1~0.45，nDCG@20~0.625），门控激活后进一步提升（Recall@1~0.56，nDCG@20~0.699，门控率45%），紧凑骨干模型亦获相似收益。编码器基线在效果和灵活性上均落后。结果表明，以默认快速排序、必要时选择生成解释，是高精度与可接受成本间的实用平衡；单策略设计简化部署与预算规划，课程训练原则亦可推广至其他决策任务。
原文摘要 · Abstract (English)
Clinicians need ranking systems that work in real time and still justify their choices. Motivated by the need for a low-latency, decoder-based reranker, we present OG-Rank, a single-decoder approach that pairs a pooled first-token scoring signal with an uncertainty-gated explanation step. The model scores all candidates in one pass and generates a brief, structured rationale only when the list is genuinely ambiguous, keeping latency predictable. Trained with a curriculum that concentrates effort on hard cases, OG-Rank delivers strong effectiveness on encounter-scoped order selection (fast path: Recall@1~0.45, nDCG@20~0.625) and improves further when the gate activates (Recall@1~0.56, nDCG@20~0.699 at a 45\% gate rate), while compact backbones show similar gains under the same policy. Encoder baselines trail in both effectiveness and flexibility. The result is a practical recipe: rank fast by default and explain when it helps, a pattern that applies broadly to decision tasks where selective generation buys accuracy at acceptable cost. The single-policy design simplifies deployment and budget planning, and the curriculum principle (spend more on the hard cases, less on the easy ones) readily transfers beyond clinical order selection.
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。