Research map / Learning theory and algorithms / Combinatorial and search optimization
Fast k-Nearest Neighbour Search via Prioritized DCI
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 |
|---|---|---|---|
| Prioritize the simple index with the closest next projected point instead of cycling through indices | DCI · CIFAR-100 | distance evaluations: 87.2% to 92.5% fewer distance evaluations | Helped |
| Prioritize the simple index with the closest next projected point instead of cycling through indices | DCI · MNIST | distance evaluations: 96.4% to 97.0% fewer distance evaluations | Helped |
| Use Prioritized DCI rather than product quantization | product quantization · CIFAR-100 | distance evaluations: 91.7% to 92.8% fewer distance evaluations | Helped |
| Use Prioritized DCI rather than product quantization | product quantization · MNIST | distance evaluations: 87.1% to 89.8% fewer distance evaluations | Helped |
| Use Prioritized DCI rather than LSH | LSH · CIFAR-100 | distance evaluations: 90.9% to 93.8% fewer distance evaluations | Helped |
| Use Prioritized DCI rather than LSH | LSH · MNIST | distance evaluations: 98.8% to 99.3% fewer distance evaluations | Helped |
| Use the space-efficient setting, m = 10 and L = 2 | Prioritized DCI · CIFAR-100 | distance evaluations vs LSH: 6-fold reduction | Helped |
| Use the space-efficient setting, m = 10 and L = 2 | Prioritized DCI · MNIST | distance evaluations vs LSH: 32-fold reduction | Helped |
| Use Prioritized DCI with standard DCI hyperparameters rather than LSH | LSH · CIFAR-100 | memory consumption: 95.5% less space; 22-fold reduction | Helped |
| Use Prioritized DCI with standard DCI hyperparameters rather than LSH | LSH · MNIST | memory consumption: 95.3% less space; 21-fold reduction | Helped |
| Use the space-efficient setting, m = 10 and L = 2, rather than LSH | LSH · CIFAR-100 | memory consumption: 98.2% less space; 55-fold reduction | Helped |
| Use the space-efficient setting, m = 10 and L = 2, rather than LSH | LSH · MNIST | memory consumption: 97.9% less space; 48-fold reduction | Helped |
Closest papers by meaning
- Fast k-Nearest Neighbour Search via Dynamic Continuous Indexing (ICML 2016 · similarity 0.94)
- Falconn++: A Locality-sensitive Filtering Approach for Approximate Nearest Neighbor Search (NeurIPS 2022 · similarity 0.88)
- Shortlist Selection With Residual-Aware Distance Estimator for K-Nearest Neighbor Search (CVPR 2016 · similarity 0.86)
- Sublinear Time Nearest Neighbor Search over Generalized Weighted Space (ICML 2019 · similarity 0.86)
- MCGI: Manifold-Consistent Graph Indexing for Billion-Scale Disk-Resident Vector Search (NeurIPS 2026 · similarity 0.85)
- Graph-Based Algorithms for Diverse Similarity Search (ICML 2025 · similarity 0.84)
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.