bibtex short conference paper to appear
Appearing in: 47th European Conference on Information Retrieval (ECIR 2025)
Abstract:
Multi-vector retrieval methods, exemplified by the ColBERT architecture, have shown substantial promise for retrieval by providing strong trade-offs in terms of retrieval latency and effectiveness. However, they come at a high cost in terms of storage since a (potentially com- pressed) vector needs to be stored for every token in the input collection. To overcome this issue, we propose encoding documents to a fixed num- ber of vectors, which are no longer necessarily tied to the input tokens. Beyond reducing the storage costs, our approach has the advantage that document representations become of a fixed size on disk, allowing for bet- ter OS paging management. Through experiments using the MSMARCO passage corpus and BEIR with the ColBERT-v2 architecture, a repre- sentative multi-vector ranking model architecture, we find that passages can be effectively encoded into a fixed number of vectors while retaining most of the original effectiveness.
BibTeX @inproceedings{macavaney:ecir2025-constbert, author = {MacAvaney, Sean and Mallia, Antonio and Tonellotto, Nicola}, title = {Efficient Constant-Space Multi-Vector Retrieval}, booktitle = {47th European Conference on Information Retrieval}, year = {2025} }