提出层次化语料编码器，让检索模型更灵活且性能更强。
Hierarchical corpus encoder: Fusing generative retrieval and dense indices
- 基于文档层级结构设计编码器，融合生成式检索与密集索引优势。
- 在零样本和有监督设置下均优于生成式检索模型。
- 支持动态增删文档，适合需要频繁更新的检索系统。
生成式检索利用序列模型根据查询条件生成文档编号（DSI (Tay et al., 2022); NCI (Wang et al., 2022) 等）。尽管这提升了零样本检索性能，但难以支持训练中未见的文档。我们发现生成式检索的性能依赖于文档层级中兄弟节点间的对比学习。由此提出层次化语料编码器（HCE），可兼容传统密集编码器。实验表明，HCE 在无监督零样本与有监督设置下均优于生成式检索模型，同时支持索引中文档的便捷增删。
原文摘要 · Abstract (English)
Generative retrieval employs sequence models for conditional generation of document IDs based on a query (DSI (Tay et al., 2022); NCI (Wang et al., 2022); inter alia). While this has led to improved performance in zero-shot retrieval, it is a challenge to support documents not seen during training. We identify the performance of generative retrieval lies in contrastive training between sibling nodes in a document hierarchy. This motivates our proposal, the hierarchical corpus encoder (HCE), which can be supported by traditional dense encoders. Our experiments show that HCE achieves superior results than generative retrieval models under both unsupervised zero-shot and supervised settings, while also allowing the easy addition and removal of documents to the index.
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。