All-in-one Graph-based Indexing For Hybrid Search On Gpus

Abstract

Hybrid search has emerged as a promising paradigm that combines lexical and semantic retrieval, enhancing accuracy for applications such as recommendations, information retrieval, and Retrieval-Augmented Generation. However, existing methods are constrained by a trilemma: they sacrifice flexibility for efficiency, suffer from accuracy degradation, or incur prohibitive storage overhead for flexible combinations of retrieval paths. This paper introduces Allan-Poe, a novel all-in-one graph index accelerated by GPUs for efficient hybrid search. We first analyze the limitations of existing retrieval paradigms and extract key design principles for an effective hybrid index. Guided by the principles, we architect a unified graph-based index that flexibly integrates three retrieval paths (dense vector, sparse vector, and full-text) within a single, cohesive structure. To enable efficient construction, we design a GPU-accelerated pipeline featuring a warp-level hybrid distance kernel, RNG-IP jo

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keyli2025all

Related papers

• DGAI: Decoupled On-disk Graph-based ANN Index For Efficient Updates And Queries (2025)0.00
• Billion-scale Similarity Search With Gpus (2017)24.96
• Hybrid Inverted Index Is A Robust Accelerator For Dense Retrieval (2022)7.07
• A Generic Inverted Index Framework For Similarity Search On The GPU - Technical Report (2016)0.00
• CAGRA: Highly Parallel Graph Construction And Approximate Nearest Neighbor Search For Gpus (2023)12.17
• Efficient And Effective Retrieval Of Dense-sparse Hybrid Vectors Using Graph-based Approximate Nearest Neighbor Search (2024)0.00
• GGNN: Graph-based GPU Nearest Neighbor Search (2019)13.39
• Gpu-accelerated Multi-relational Parallel Graph Retrieval For Web-scale Recommendations (2025)0.00