用主动学习提升大模型排序效率，减少调用次数同时提高排名质量。
Active Learners as Efficient PRP Rerankers
- 将排序任务转为有噪声的主动学习，按需选择关键比较对
- 相同调用次数下，NDCG@10 提升显著，优于传统排序方法
- 单次调用即可完成方向随机化，消除位置偏差，适合资源受限场景
成对排序提示（PRP）通过大语言模型获取成对偏好判断，并用经典排序算法聚合为最终排名。然而，这些判断存在噪声、顺序敏感和非传递性问题，导致排序假设不成立。由于排序旨在恢复完整排列，截断至前K项难以保证可靠性。因此，我们将PRP重排序重构为从有噪声成对比较中进行主动学习，证明主动排序器可作为即插即用替代方案，在调用受限条件下提升每调用一次的NDCG@10表现。我们的抗噪框架引入了随机方向的查询代理，每对比较仅需一次大模型调用，将系统性位置偏差转化为零均值噪声，从而实现无偏聚合，且无需双向调用开销。
原文摘要 · Abstract (English)
Pairwise Ranking Prompting (PRP) elicits pairwise preference judgments from an LLM, which are then aggregated into a ranking, usually via classical sorting algorithms. However, judgments are noisy, order-sensitive, and sometimes intransitive, so sorting assumptions do not match the setting. Because sorting aims to recover a full permutation, truncating it to meet a call budget does not produce a dependable top-K. We thus reframe PRP reranking as active learning from noisy pairwise comparisons and show that active rankers are drop-in replacements that improve NDCG@10 per call in the call-constrained regime. Our noise-robust framework also introduces a randomized-direction oracle that uses a single LLM call per pair. This approach converts systematic position bias into zero-mean noise, enabling unbiased aggregate ranking without the cost of bidirectional calls.
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。