Pith. sign in

End-to-End Retrieval with Learned Dense and Sparse Representations Using Lucene

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it
abstract

The bi-encoder architecture provides a framework for understanding machine-learned retrieval models based on dense and sparse vector representations. Although these representations capture parametric realizations of the same underlying conceptual framework, their respective implementations of top-$k$ similarity search require the coordination of different software components (e.g., inverted indexes, HNSW indexes, and toolkits for neural inference), often knitted together in complex architectures. In this work, we ask the following question: What's the simplest design, in terms of requiring the fewest changes to existing infrastructure, that can support end-to-end retrieval with modern dense and sparse representations? The answer appears to be that Lucene is sufficient, as we demonstrate in Anserini, a toolkit for reproducible information retrieval research. That is, effective retrieval with modern single-vector neural models can be efficiently performed directly in Java on the CPU. We examine the implications of this design for information retrieval researchers pushing the state of the art as well as for software engineers building production search systems.

citation-role summary

background 1

citation-polarity summary

fields

cs.CY 1

years

2025 1

verdicts

UNVERDICTED 1

roles

background 1

polarities

unclear 1

representative citing papers

citing papers explorer

Showing 1 of 1 citing paper.