卓越
目录平台模板

Multi-Vector Retrieval Algorithm that reduces multi-vector similarity search to single-vector similarity search via Fixed Dimensional Encodings. Achieves 10% improved recall with 90% lower latency compared to existing approaches.

A commutative vector quantization method for KV cache compression that reduces FP16 cache size by 87.5% with 2-bit quantization and enables 1-bit quantization, allowing LLaMA-3.1 8B to run with 128K context on a single RTX 4090 GPU.

Research paper introducing CAPS, a practical partition index designed for filtered similarity search. Published as an arXiv preprint in 2023 by Gaurav Gupta et al., it addresses the challenge of combining attribute filtering with approximate nearest neighbor search efficiently.

An efficient CPU/GPU cooperative processing architecture for billion-scale approximate nearest neighbor search. FusionANNS achieves up to 13.1× higher QPS compared to SPANN and can handle billion-vector datasets with over 12,000 QPS while maintaining 15ms latency using only one entry-level GPU.

SPANN is a highly efficient billion-scale ANN search system using clustered HNSW indexes with dynamic partitioning for balanced load. Key features: disk-based, high recall, low latency on commodity hardware. Use cases: web-scale recommendation, image retrieval. Improves on DiskANN with better build time; competitive FAISS GPU in CPU perf.

BatANN system by Dang et al. for high-throughput distributed disk-based vector search. Supports scalable ANN in distributed environments.

CoTra system by Zhi et al. for efficient distributed vector search using RDMA. Published in SIGMOD 2026 proceedings.

Research paper by Chen et al. examining the meaningfulness of nearest neighbor search in high-dimensional spaces. Analyzes limitations and implications for vector similarity search. Key for understanding ANN effectiveness.

FusionANNS architecture by Bing Tian et al. for billion-scale ANN search using CPU/GPU cooperation.

Paper by Tepper et al. proposing GleanVec, a method to accelerate vector search using minimalist nonlinear dimensionality reduction. Improves efficiency for high-dimensional vector queries.

iDEC by Gong et al. for approximate nearest neighbor search using indexable distance estimating codes. VLDB Endowment 13.9 (2020).

Highly-efficient billion-scale approximate nearest neighbor search algorithm introduced by Chen et al. Focuses on scalability and performance for large datasets in high-dimensional spaces. Relevant for vector database indexing techniques.