让检索模型动态学习新实体知识，提升冷门实体查询准确率。
Dynamic Injection of Entity Knowledge into Dense Retrievers
- 通过上下文-实体注意力机制和可更新的实体嵌入增强BERT检索器
- 在三个数据集上显著提升检索精度，尤其在EntityQuestions上增益明显
- 无需重训练即可注入外部知识，适合需要快速适应新领域的应用
密集检索器常因实体知识有限，在涉及低频实体的查询上表现不佳。本文提出知识增强型文档检索器（KPR），基于BERT架构，引入上下文-实体注意力层与可动态更新的实体嵌入。该设计使KPR可在不重新训练的前提下融入外部实体知识。在三个数据集上的实验表明，KPR持续提升检索准确率，尤其在EntityQuestions数据集上表现突出。以现成的bge-base检索器为基础，KPR在两个数据集上达到同规模模型的最先进水平。模型与代码已开源至https://github.com/knowledgeable-embedding/knowledgeable-embedding。
原文摘要 · Abstract (English)
Dense retrievers often struggle with queries involving less-frequent entities due to their limited entity knowledge. We propose the Knowledgeable Passage Retriever (KPR), a BERT-based retriever enhanced with a context-entity attention layer and dynamically updatable entity embeddings. This design enables KPR to incorporate external entity knowledge without retraining. Experiments on three datasets demonstrate that KPR consistently improves retrieval accuracy, with particularly large gains on the EntityQuestions dataset. When built on the off-the-shelf bge-base retriever, KPR achieves state-of-the-art performance among similarly sized models on two datasets. Models and code are released at https://github.com/knowledgeable-embedding/knowledgeable-embedding.
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。