Accelerating Graph-based ANNS with Adaptive Awareness

ICML 2019 paper introducing a learned routing approach for similarity graphs, using machine learning to guide greedy search traversal in graph-based approximate nearest neighbor search.

Foundational paper introducing the navigable small world (NSW) graph algorithm for approximate nearest neighbor search, which became the basis for widely-used graph-based ANN methods including HNSW.

Paper from 2024 proposing a probabilistic routing approach for graph-based approximate nearest neighbor search, introducing probability models to guide search traversal on proximity graphs.

VLDB 2024 paper proposing RoarGraph, a projected bipartite graph structure for efficient cross-modal approximate nearest neighbor search. The method addresses the challenges of searching across different modalities (e.g., text, image) using graph-based indexing.

iRangeGraph is an ANN indexing approach and accompanying implementation for range-filtering nearest neighbor search. It provides a specialized graph-based index that supports vector similarity search under range constraints, making it directly useful as a component or reference implementation for advanced vector database indexing and retrieval.

NSG is an approximate nearest neighbor search algorithm based on a sparse navigable graph structure designed for high-dimensional vector similarity search. The reference implementation provides a graph-based ANN index that can be integrated into custom vector retrieval systems.