Document Expansion by Query Prediction
Rodrigo Nogueira,1 Wei Yang,2 Jimmy Lin,2 Kyunghyun Cho3,4,5,6 1 Tandon School of Engineering, New York University 2 David R. Cheriton School of Computer Science, University of Waterloo 3 Courant Institute of Mathematical Sciences, New York University 4 Center for Data Science, New York University 5 Facebook AI Research 6 CIFAR Azrieli Global Scholar
Abstract
One technique to improve the retrieval effectiveness of a search engine is to expand documents with terms that are related or representative of the documents’ content. From the perspective of a question answering system, this might comprise questions the document can potentially answer. Following this observation, we propose a simple method that predicts which queries will be issued for a given document and then expands it with those predictions with a vanilla sequence-to-sequence model, trained using datasets consisting of pairs of query and relevant documents. By combining our method with a highly-effective re-ranking component, we achieve the state of the art in two retrieval tasks. In a latency-critical regime, retrieval results alone (without re-ranking) approach the effectiveness of more computationally expensive neural re-rankers but are much faster.
中文速览
搜索引擎面临"词汇不匹配"难题——用户的查询词往往与文档中的表述不同，导致相关文档被漏掉。研究者提出了一种叫做 Doc2query 的文档扩展方法：先用序列到序列（sequence-to-sequence）的 Transformer 模型，学习"给定文档、预测用户可能提问什么"，再把生成的预测查询拼接到原文档里一起建索引，让传统的 BM25 检索能直接受益于神经网络的语义理解，而无需在检索时实时运行昂贵的神经推理。实验在 MS MARCO 和 TREC-CAR 两个基准上表明，文档扩展使检索效果提升约 15%，进一步结合 BERT 重排序后达到当时最优水平，且在对延迟敏感的场景下，仅用扩展后的 BM25 检索（不加重排序）就比同类神经重排序器快约七倍。这项工作的重要意义在于，它首次证明了神经网络驱动的文档扩展切实可行，为将神经网络的计算开销从在线检索转移到离线索引阶段提供了一条简单且可复现的路径。
原文 arXiv:1904.08375;中英对照 + 大白话阅读 https://aha.fim.ai/paper/1904.08375v2