# DyVo: Dynamic Vocabularies for Learned Sparse Retrieval with Entities

arXiv: https://arxiv.org/abs/2410.07722 · HTML: https://mlsanity.com/paper/2410.07722/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 10 Oct 2024.

## Built on by (most cited first, 8 of 8)

- [Enhancing Lexicon-Based Text Embeddings with Large Language Models](https://mlsanity.com/paper/2501.09749/) (2025)
- [Entity Labels Are Not Entity Signals: A Framework for Observable Relevance in Document Re-Ranking](https://mlsanity.com/paper/2606.15998/) (2026)
- [Dynamic Injection of Entity Knowledge into Dense Retrievers](https://mlsanity.com/paper/2507.03922/) (2025)
- [REGENT: Relevance-Guided Attention for Entity-Aware Multi-Vector Neural Re-Ranking](https://mlsanity.com/paper/2510.11592/) (2025)
- [To Case or Not to Case: An Empirical Study in Learned Sparse Retrieval](https://mlsanity.com/paper/2601.17500/) (2026)
- [From Tokens to Concepts: Leveraging SAE for SPLADE](https://mlsanity.com/paper/2604.21511/) (2026)
- [Robustness of IR Models to Collection Growth](https://mlsanity.com/paper/2608.23419/) (2026)
- [Mitigating Entity Type Confusion in Cross-Domain NER via Multidimensional Quantification and Reasoning Enhancement](https://mlsanity.com/paper/2609.24357/) (2026)

## Abstract (arXiv)

Learned Sparse Retrieval (LSR) models use vocabularies from pre-trained transformers, which often split entities into nonsensical fragments. Splitting entities can reduce retrieval accuracy and limits the model's ability to incorporate up-to-date world knowledge not included in the training data. In this work, we enhance the LSR vocabulary with Wikipedia concepts and entities, enabling the model to resolve ambiguities more effectively and stay current with evolving knowledge. Central to our approach is a Dynamic Vocabulary (DyVo) head, which leverages existing entity embeddings and an entity retrieval component that identifies entities relevant to a query or document. We use the DyVo head to generate entity weights, which are then merged with word piece weights to create joint representations for efficient indexing and retrieval using an inverted index. In experiments across three entity-rich document ranking datasets, the resulting DyVo model substantially outperforms state-of-the-art baselines.