# ColBERTSaR: Sparsified ColBERT Index via Product Quantization

arXiv: https://arxiv.org/abs/2606.05568 · HTML: https://mlsanity.com/paper/2606.05568/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 23 of its 32 references that are in the map.
- First posted on arXiv 4 Jun 2026.

## Builds on (most cited first, 23 of 23)

- [Dense Passage Retrieval for Open-Domain Question Answering](https://mlsanity.com/paper/2004.04906/) (2020)
- [Billion-scale similarity search with GPUs](https://mlsanity.com/paper/1702.08734/) (2017)
- [Qwen3 Embedding: Advancing Text Embedding and Reranking Through Foundation Models](https://mlsanity.com/paper/2506.05176/) (2025)
- [BEIR: A Heterogenous Benchmark for Zero-shot Evaluation of Information Retrieval Models](https://mlsanity.com/paper/2104.08663/) (2021)
- [ColBERT: Efficient and Effective Passage Search via Contextualized Late Interaction over BERT](https://mlsanity.com/paper/2004.12832/) (2020)
- [ColBERTv2: Effective and Efficient Retrieval via Lightweight Late Interaction](https://mlsanity.com/paper/2112.01488/) (2021)
- [SPLADE: Sparse Lexical and Expansion Model for First Stage Ranking](https://mlsanity.com/paper/2107.05720/) (2021)
- [SPLADE v2: Sparse Lexical and Expansion Model for Information Retrieval](https://mlsanity.com/paper/2109.10086/) (2021)
- [PLAID: An Efficient Engine for Late Interaction Retrieval](https://mlsanity.com/paper/2205.09707/) (2022)
- [SPLADE-v3: New baselines for SPLADE](https://mlsanity.com/paper/2403.06789/) (2024)
- [Learning Passage Impacts for Inverted Indexes](https://mlsanity.com/paper/2104.12016/) (2021)
- [Rethinking the Role of Token Retrieval in Multi-Vector Retrieval](https://mlsanity.com/paper/2304.01982/) (2023)
- [Translate-Distill: Learning Cross-Language Dense Retrieval by Translation and Distillation](https://mlsanity.com/paper/2401.04810/) (2024)
- [CITADEL: Conditional Token Interaction via Dynamic Lexical Routing for Efficient and Effective Multi-Vector Retrieval](https://mlsanity.com/paper/2211.10411/) (2022)
- [A Unified Framework for Learned Sparse Retrieval](https://mlsanity.com/paper/2303.13416/) (2023)
- [WARP: An Efficient Engine for Multi-Vector Retrieval](https://mlsanity.com/paper/2501.17788/) (2025)
- [JaColBERTv2.5: Optimising Multi-Vector Retrievers to Create State-of-the-Art Japanese Retrievers with Constrained Resources](https://mlsanity.com/paper/2407.20750/) (2024)
- [A White Box Analysis of ColBERT](https://mlsanity.com/paper/2012.09650/) (2020)
- [A Reproducibility Study of PLAID](https://mlsanity.com/paper/2404.14989/) (2024)
- [NeuCLIRBench: A Modern Evaluation Collection for Monolingual, Cross-Language, and Multilingual Information Retrieval](https://mlsanity.com/paper/2511.14758/) (2025)
- [SPLATE: Sparse Late Interaction Retrieval](https://mlsanity.com/paper/2404.13950/) (2024)
- [Distillation for Multilingual Information Retrieval](https://mlsanity.com/paper/2405.00977/) (2024)
- [NeuCLIRTech: Chinese Monolingual and Cross-Language Information Retrieval Evaluation in a Challenging Domain](https://mlsanity.com/paper/2602.05334/) (2026)

## Abstract (arXiv)

While ColBERT is an effective neural retrieval architecture, it requires a heavy index structure to support candidate set retrieval based on approximated token embeddings, gathering and decompressing document token embeddings, and applying the MaxSim operation. Indexes in PLAID and similar ColBERT implementations require five to ten times the disk storage of the original raw text, which limits their scalability. Furthermore, prior work has identified that the gathering and decompression stages are the primary inefficiencies at query time. Limiting the number of document tokens that must be gathered by thresholding and score approximation does not eliminate the need for the entire index to support ad hoc queries. In this work, we propose an embedding quantization approach that turns a ColBERT index into a true inverted index. We show that, theoretically, ColBERT with embedding quantization is equivalent to learned-sparse retrieval except for the scoring mechanism. Empirically, we demonstrate that our index is 50-70% smaller than a one-bit PLAID index while retaining retrieval effectiveness.