Approximate Nearest Neighbor Search and Lightweight Dense Vector Reranking in Multi-Stage Retrieval Architectures
By Zhengkai Tu, Wei Yang, Zihang Fu, Yuqing Xie, Luchen Tan, Kun Xiong, Ming Li, Jimmy J. Lin. 2020
17 citations recorded by the source corpus.
In the context of a multi-stage retrieval architecture, we explore candidate generation based on approximate nearest neighbor (ANN) search and lightweight reranking based on dense vector representations. These results serve as input to slower but more accurate rerankers such as those based on transformers. Our goal is to characterize the effectiveness-efficiency tradeoff space in this context. We find that, on sentence-length segments of text, ANN techniques coupled with dense vector reranking dominate approaches based on inverted indexes, and thus our proposed design should be preferred. For paragraph-length segments, ANN-based and index-based techniques share the Pareto frontier, which means that the choice of alternatives depends on the desired operating point.
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
- C. Lassance
- Thibault Formal
- Benjamin Piwowarski
- Stéphane Clinchant
- Niklas Muennighoff
- Keshav Santhanam
- Omar Khattab
Citations in the corpus
- SPLADE: Sparse Lexical and Expansion Model for First Stage Ranking (2021, unknown): “To the best of our knowledge, very few studies have discussed the impact of using approximate nearest neighbors (ANN) search on IR metrics [2, 23].”
- Autonomous chemical research with large language models (2023, unknown): “Information retrieval systems are usually based on two candidate selection approaches: inverted search index and vector database 38–41 .”
- SPLADE v2: Sparse Lexical and Expansion Model for Information Retrieval (2021, background): “To the best of our knowledge, very few studies have discussed the impact of using approximate nearest neighbors (ANN) search on IR metrics [2, 27].”