Hybrid-Vector Retrieval for Visually Rich Documents: Combining Single-Vector Efficiency and Multi-Vector Accuracy

TL;DR

HEAVEN is a hybrid-vector retrieval framework that combines the efficiency of single-vector methods with the accuracy of multi-vector approaches for visually rich document retrieval.

Contribution

The paper introduces HEAVEN, a novel two-stage hybrid-vector framework that significantly improves retrieval efficiency while maintaining high accuracy for visually rich documents.

Findings

HEAVEN achieves 99.87% of multi-vector recall@1 performance.

Reduces per-query computation by 99.82%.

Introduces ViMDoc benchmark for realistic document retrieval evaluation.

Abstract

Retrieval over visually rich documents is essential for tasks such as legal discovery, scientific search, and enterprise knowledge management. Existing approaches fall into two paradigms: single-vector retrieval, which is efficient but coarse, and multi-vector retrieval, which is accurate but computationally expensive. To address this trade-off, we propose HEAVEN, a plug-and-play two-stage hybrid-vector framework. In the first stage, HEAVEN efficiently retrieves candidate pages using a single-vector method over Visually-Summarized Pages (VS-Pages), which assemble representative visual layouts from multiple pages. In the second stage, it reranks candidates with a multi-vector method while filtering query tokens by linguistic importance to reduce redundant computations. To evaluate retrieval systems under realistic conditions, we also introduce ViMDoc, a benchmark for visually rich, multi-document, and long-document retrieval. Across four benchmarks, HEAVEN attains 99.87% of the Recall@1 performance of multi-vector models on average while reducing per-query computation by 99.82%, achieving efficiency and accuracy. Our code and datasets are available at: https://github.com/juyeonnn/HEAVEN