Representation Sparsification with Hybrid Thresholding for Fast SPLADE-based Document Retrieval

Representation Sparsification with Hybrid Thresholding for Fast SPLADE-based Document Retrieval
复制标题

DOI:
10.1145/3539618.3592051
复制
发表时间:
2023-06
期刊:
Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval
影响因子:
--
通讯作者:
Yifan Qiao;Yingrui Yang;Shanxiu He;Tao Yang
Yifan Qiao;Yingrui Yang;Shanxiu He;Tao Yang
中科院分区:
其他
文献类型:
--
作者:
Yifan Qiao;Yingrui Yang;Shanxiu He;Tao Yang

文献摘要

相似文献

使用基于变压器的神经模型学习稀疏文档表示在相关性有效性和时间效率方面都很有吸引力。本文提出了一种基于软硬阈值和倒排索引近似的表示稀疏化方案,以实现更快的基于splade的文档检索。给出了这种可学习混合阈值方案的影响分析和实验结果。
Learned sparse document representations using a transformer-based neural model has been found to be attractive in both relevance effectiveness and time efficiency. This paper describes a representation sparsification scheme based on hard and soft thresholding with an inverted index approximation for faster SPLADE-based document retrieval. It provides analytical and experimental results on the impact of this learnable hybrid thresholding scheme.