Research map / Learning theory and algorithms / Prediction and decision losses
Asymmetric Deep Supervised Hashing
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 |
|---|---|---|---|
| Use asymmetric deep hashing with directly learned database codes instead of COSDISH-D | ADSH · CIFAR-10, 12 bits | MAP: 0.8431 -> 0.8466 | Helped |
| Use asymmetric deep hashing with directly learned database codes instead of COSDISH-D | ADSH · CIFAR-10, 24 bits | MAP: 0.8666 -> 0.9062 | Helped |
| Use asymmetric deep hashing with directly learned database codes instead of COSDISH-D | ADSH · CIFAR-10, 32 bits | MAP: 0.8738 -> 0.9175 | Helped |
| Use asymmetric deep hashing with directly learned database codes instead of COSDISH-D | ADSH · CIFAR-10, 48 bits | MAP: 0.8724 -> 0.9263 | Helped |
| Use asymmetric deep hashing with directly learned database codes instead of SDH-D | ADSH · NUS-WIDE, 12 bits | MAP: 0.8478 -> 0.8570 | Helped |
| Use asymmetric deep hashing with directly learned database codes instead of SDH-D | ADSH · NUS-WIDE, 24 bits | MAP: 0.8683 -> 0.8939 | Helped |
| Use asymmetric deep hashing with directly learned database codes instead of SDH-D | ADSH · NUS-WIDE, 32 bits | MAP: 0.8647 -> 0.9008 | Helped |
| Use asymmetric deep hashing with directly learned database codes instead of LFH-D | ADSH · NUS-WIDE, 48 bits | MAP: 0.8721 -> 0.9074 | Helped |
| Directly learn database binary codes rather than generate them with the learned hash function | Lin:V, LFH-D, SDH-D, COSDISH-D · CIFAR-10 and NUS-WIDE | MAP | Helped |
| Increase the number of sampled query points | ADSH · CIFAR-10 | MAP | Helped |
| Vary gamma within the tested range | ADSH · CIFAR-10 | MAP | No clear effect |
| Use CNN-F rather than AlexNet for DHN | DHN · CIFAR-10 | MAP: 0.5512 -> 0.5978; 0.6091 -> 0.6399; 0.6114 -> 0.6471; 0.6433 -> 0.6870 | Helped |
Closest papers by meaning
- AMVH: Asymmetric Multi-Valued Hashing (CVPR 2017 · similarity 0.87)
- SAH: Shifting-Aware Asymmetric Hashing for Reverse k Maximum Inner Product Search (AAAI 2023 · similarity 0.87)
- Fast Online Hashing with Multi-Label Projection (AAAI 2023 · similarity 0.87)
- DistillHash: Unsupervised Deep Hashing by Distilling Data Pairs (CVPR 2019 · similarity 0.86)
- Asymmetric Discrete Graph Hashing (AAAI 2017 · similarity 0.86)
- Pairwise-Label-Based Deep Incremental Hashing with Simultaneous Code Expansion (AAAI 2024 · similarity 0.85)
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.