Research map / Graph learning / Graph algorithms and clustering
Link and Code: Fast Indexing With Graphs and Compact Regression Codes
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 |
|---|---|---|---|
| combine graph traversal with compact vector codes and quantized neighbor-regression codes | L&C · BIGANN | recall@1: 0.330 -> 0.461 | Helped |
| combine graph traversal with compact vector codes and quantized neighbor-regression codes | L&C · Deep1B | recall@1: 0.456 -> 0.668 | Helped |
| use graph-based search rather than IMI | nearest-neighbor search · Deep1M | selectivity at a desired accuracy: 5 to 8 times | Helped |
| add quantized regression coefficients for neighbor-based refinement (ablation) | L6&OPQ32 M=8 · Deep1B | recall@1, T=16384: 0.434 -> 0.438 | Helped |
| add quantized regression coefficients for neighbor-based refinement (ablation) | L6&OPQ32 M=8 · Deep1B | vector quantization error (× 10^3): 24.3 -> 19.8 | Helped |
| replace OPQ40 with OPQ40 M=0 shared-weight refinement (ablation) | L6&OPQ40 · Deep1B | recall@1, T=16384: 0.434 -> 0.435 | No clear effect |
| add 0-byte shared-weight refinement using graph neighbors (ablation) | L&C · Deep1B | recall@1 | No clear effect |
| use a two-level residual codec rather than a one-level codec (ablation) | vector encoding · Deep1M | recall@1: 0.604 -> 0.731 | Helped |
| use OPQ rather than PCA dimensionality reduction (ablation) | vector encoding · Deep1M | recall@1: 0.017 -> 0.604 | Helped |
| allocate more of a fixed memory budget to regression codes and less to the initial OPQ code (ablation) | L6&C · Deep1B | recall@1, T=16384: 0.434 -> 0.438 | Helped |
Closest papers by meaning
- Explore-Exploit Graph Traversal for Image Retrieval (CVPR 2019 · similarity 0.89)
- Efficient Large-Scale Similarity Search Using Matrix Factorization (CVPR 2016 · similarity 0.86)
- Supervised Quantization for Similarity Search (CVPR 2016 · similarity 0.86)
- Guided Similarity Separation for Image Retrieval (NeurIPS 2019 · similarity 0.86)
- Graph-Based Algorithms for Diverse Similarity Search (ICML 2025 · similarity 0.84)
- Exploiting Distribution Constraints for Scalable and Efficient Image Retrieval (ICLR 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.