Billion-scale Similarity Search Using a Hybrid Indexing Approach with Advanced Filtering

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Emanuilov, Simeon, Dimov, Aleksandar
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866915117784891392
author Emanuilov, Simeon
Dimov, Aleksandar
author_facet Emanuilov, Simeon
Dimov, Aleksandar
contents This paper presents a novel approach for similarity search with complex filtering capabilities on billion-scale datasets, optimized for CPU inference. Our method extends the classical IVF-Flat index structure to integrate multi-dimensional filters. The proposed algorithm combines dense embeddings with discrete filtering attributes, enabling fast retrieval in high-dimensional spaces. Designed specifically for CPU-based systems, our disk-based approach offers a cost-effective solution for large-scale similarity search. We demonstrate the effectiveness of our method through a case study, showcasing its potential for various practical uses.
format Preprint
id arxiv_https___arxiv_org_abs_2501_13442
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Billion-scale Similarity Search Using a Hybrid Indexing Approach with Advanced Filtering
Emanuilov, Simeon
Dimov, Aleksandar
Information Retrieval
Databases
Distributed, Parallel, and Cluster Computing
Machine Learning
68P20, 62H30, 68T07
H.3.3; I.2.6; G.1.0
This paper presents a novel approach for similarity search with complex filtering capabilities on billion-scale datasets, optimized for CPU inference. Our method extends the classical IVF-Flat index structure to integrate multi-dimensional filters. The proposed algorithm combines dense embeddings with discrete filtering attributes, enabling fast retrieval in high-dimensional spaces. Designed specifically for CPU-based systems, our disk-based approach offers a cost-effective solution for large-scale similarity search. We demonstrate the effectiveness of our method through a case study, showcasing its potential for various practical uses.
title Billion-scale Similarity Search Using a Hybrid Indexing Approach with Advanced Filtering
topic Information Retrieval
Databases
Distributed, Parallel, and Cluster Computing
Machine Learning
68P20, 62H30, 68T07
H.3.3; I.2.6; G.1.0
url https://arxiv.org/abs/2501.13442