Research map / Label-efficient and robust learning / Recognition and long-tail
DistillHash: Unsupervised Deep Hashing by Distilling Data Pairs
Check 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 |
|---|---|---|---|
| use DistillHash rather than SSDH | unsupervised deep hashing · FLICKR25K | MAP: 3.08% average MAP improvement | Helped |
| use DistillHash rather than SSDH | unsupervised deep hashing · NUSWIDE | MAP: 4.01% average MAP improvement | Helped |
| use DistillHash rather than SSDH | unsupervised deep hashing · CIFAR10 | MAP: 2.86% average MAP improvement | Helped |
| add distilled-pair training instead of using initial noisy similarity labels (ablation) | DistillHash Bayesian learning framework · FLICKR25K | MAP: 3.11%, 4.23%, 3.49% and 2.11% at 16, 32, 64, and 128 bits | Helped |
| add distilled-pair training instead of using initial noisy similarity labels (ablation) | DistillHash Bayesian learning framework · NUSWIDE | MAP: 3.45%, 3.95%, 2.89% and 2.96% at 16, 32, 64, and 128 bits | Helped |
| add distilled-pair training instead of using initial noisy similarity labels (ablation) | DistillHash Bayesian learning framework · CIFAR10 | MAP: 2.97%, 3.15%, 2.94% and 3.12% at 16, 32, 64, and 128 bits | Helped |
| vary nearest-neighbor parameter o from 2 to 20 (ablation) | DistillHash · NUSWIDE | MAP | No clear effect |
| vary probability-layer dimension p from 32 to 128 (ablation) | DistillHash · NUSWIDE | MAP | No clear effect |
Closest papers by meaning
- Deep Unsupervised Hashing via External Guidance (ICML 2025 · similarity 0.89)
- KALAHash: Knowledge-Anchored Low-Resource Adaptation for Deep Hashing (AAAI 2025 · similarity 0.88)
- Bayesian Supervised Hashing (CVPR 2017 · similarity 0.88)
- Statistical Model-driven Similarity Hashing: Bridging Modalities for Efficient Unsupervised Retrieval (AAAI 2025 · similarity 0.88)
- Deep Supervised Discrete Hashing (NeurIPS 2017 · similarity 0.88)
- Effective Comparative Prototype Hashing for Unsupervised Domain Adaptation (AAAI 2024 · similarity 0.86)
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.