Research map / Graph learning / Graph algorithms and clustering
Graph-based Nearest Neighbor Search: From Practice to Theory
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add distance-weighted long-range edges with probability proportional to distance^{-d} | Graph-based NNS in the dense regime · Dense regime | search steps: O(\log^2 n) | Helped |
| Add random shortcut edges and choose the closest sampled node as the starting point | Graph-based NNS · Dense regime | time complexity: \Omega\left(d^{1/2}\cdot e^{\omega(1+o(1))} \cdot M^{d}\right) | No clear effect |
| Use beam search to explore the query's L-neighborhood | Graph-based NNS · Dense regime | search success probability: 1 - o(1) | Helped |
| Increase the number of long-range shortcut edges beyond about log n (ablation) | Graph-based NNS · Additional experiments | algorithm quality | No clear effect |
| Use pre-sampling of sqrt(n) nodes when adding shortcut edges (ablation) | Graph-based NNS · Uniform synthetic data | quality | Hurt |
| Use rank-based rather than distance-based shortcut-edge probabilities (ablation) | Graph-based NNS · Uniform synthetic data | quality | Mixed |
| Connect each node to a fixed number of nearest neighbors instead of nodes within a distance threshold (ablation) | Graph-based NNS · Uniform synthetic data | performance | No clear effect |
Closest papers by meaning
- Graph-Based Algorithms for Diverse Similarity Search (ICML 2025 · similarity 0.91)
- Navigable Graphs for High-Dimensional Nearest Neighbor Search: Constructions and Limits (NeurIPS 2024 · similarity 0.89)
- Probabilistic Routing for Graph-Based Approximate Nearest Neighbor Search (ICML 2024 · similarity 0.89)
- Distance Adaptive Beam Search for Provably Accurate Graph-Based Nearest Neighbor Search (NeurIPS 2025 · similarity 0.88)
- Sort Before You Prune: Improved Worst-Case Guarantees of the DiskANN Family of Graphs (ICML 2025 · similarity 0.87)
- CSPG: Crossing Sparse Proximity Graphs for Approximate Nearest Neighbor Search (NeurIPS 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.