Hybrid-Vector Retrieval for Visually Rich Documents: Combining Single-Vector Efficiency and Multi-Vector Accuracy
This addresses retrieval challenges for tasks like legal discovery and scientific search, offering a practical solution to a known bottleneck with incremental improvements.
The paper tackles the trade-off between efficiency and accuracy in retrieving visually rich documents by proposing HEAVEN, a two-stage hybrid-vector framework that achieves 99.87% of the Recall@1 performance of multi-vector models while reducing per-query computation by 99.82%.
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 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, the first 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