Document Expansion by Query Prediction
About
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.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Passage retrieval | MsMARCO (dev) | MRR@1021.5 | 116 | |
| Document Ranking | TREC DL Track 2019 (test) | nDCG@1056.9 | 96 | |
| Retrieval | MS MARCO (dev) | MRR@100.215 | 84 | |
| Passage Ranking | MS MARCO (dev) | MRR@1037.5 | 73 | |
| Passage Ranking | TREC DL 2019 (test) | NDCG@1059 | 33 | |
| Tool Retrieval | TOOLRET In-Domain (Avg) | nDCG@1051.5 | 15 | |
| Tool Retrieval | TOOLRET Zero-Shot Code | nDCG@1025.4 | 15 | |
| Tool Retrieval | TOOLRET Zero-Shot Web* | nDCG@1026.6 | 15 | |
| Tool Retrieval | TOOLRET Zero-Shot Macro-Avg | nDCG@1028.4 | 15 | |
| Tool Retrieval | TOOLRET Zero-Shot Custom | nDCG@1033.3 | 15 |