Lion-SP：软Top-K与逐项阈值优化检索效率

1. 12:00
   arXiv · Information Retrieval

   Lion-SP：用于高效检索的可学习 Soft Top-K 与逐项阈值自适应稀疏优化

   该论文提出通过可学习 soft top-K、逐项阈值（per-term thresholding）与 FLOPs 正则化协同优化模型稀疏度，以缩短 LLM 语义词扩展带来的过长查询与文档向量。在 MS MARCO 和 BEIR 数据集上，Lion-SP 显著降低平均查询与文档长度，实现更低检索延迟与存储成本，同时保持有竞争力的相关性，论文已被 SIGIR 2026 接收。

Current heat 7·Comparable peak 10（Oct 5）·Comparable change over 24 hours –

The trend compares only the same participants observed continuously; its range may be smaller than the current heat count. Move or click on the chart to inspect hourly heat; use the left and right arrow keys to switch.