TurboEmbed: An Idea About Zero-Loss Angular Quantization for High-Performance Vector Similarity Search

Read full text Summarise with VIRAT
Authors: Huili Wang
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

This paper introduces "TurboEmbed," a data-oblivious quantization framework designed to accelerate high-dimensional vector similarity searches in RAG and LLM systems. By leveraging the 360-degree angular normalization from the TurboQuant (2026) algorithm and QJL projections, TurboEmbed converts traditional floating-point cosine similarity into high-speed bitwise operations. Our implementation demonstrates a 6x-32x memory reduction and a 12x-20x theoretical throughput speedup with a correlation of >0.83 compared to FP32 baselines, enabling massive vector databases to run on consumer-grade hardware.

Fields of Study & Subjects

Advanced Image and Video Retrieval Techniques Topic Modeling Advanced Neural Network Applications

Publication details

Journal / source
Zenodo (CERN European Organization for Nuclear Research)
Publication date
27 March 2026
Language
EN
Document type
Preprint