# LoRANN

> Low-Rank Matrix Factorization algorithm for Approximate Nearest Neighbor Search, offering competitive performance with faster query times than leading libraries at various recall levels.

- **URL:** https://arxiv.org/html/2410.18926v1
- **Category:** Research Papers & Surveys
- **Tags:** Ann, algorithm, optimization
- **Updated:** 2026-03-10 11:58
- **Canonical page:** https://vectordb.ever.works/items/lorann

## Details

## Overview

LoRANN (Low-Rank Matrix Factorization for Approximate Nearest Neighbor Search) is a novel algorithm that applies matrix factorization techniques to improve ANN search efficiency.

## Key Innovation

- Low-rank matrix factorization approach
- Competitive with state-of-the-art methods
- Faster query times at various recall levels
- Efficient representation learning

## Performance

- Faster than GLASS at recall levels under 90% on most datasets
- Competitive accuracy across benchmarks
- Efficient memory utilization
- Scalable implementation

## Technical Approach

- Matrix factorization for dimension reduction
- Approximate search optimization
- Balanced accuracy-speed tradeoff
- Novel indexing strategy

## Use Cases

- High-dimensional vector search
- Large-scale similarity retrieval
- Performance-critical applications
- Research and algorithm development

## Research Context

Published in 2024, LoRANN represents recent advances in approximate nearest neighbor search algorithms, demonstrating that matrix factorization techniques can be effectively applied to vector search problems.

## Availability

Research paper available on arXiv

---

_This Markdown mirror is generated by Ever Works for AI agents. The canonical HTML page is at https://vectordb.ever.works/items/lorann._