Research map / Architectures and efficiency / RNNs and normalization
Fried Binary Embedding for High-Dimensional Visual Features
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 |
|---|---|---|---|
| Optimize the Fastfood transform's S, G, and B matrices instead of using randomly generated matrices (ablation) | Fastfood binary embedding · Holidays+1M, 1024 bits | mAP: 44.7% -> 45.6% | Helped |
| Optimize the Fastfood transform's S, G, and B matrices instead of using randomly generated matrices (ablation) | Fastfood binary embedding · Holidays+1M, 4096 bits | mAP: 46.2% -> 47.2% | Helped |
| Optimize the Fastfood transform's S, G, and B matrices instead of using randomly generated matrices (ablation) | Fastfood binary embedding · Holidays+1M, 8192 bits | mAP: 47.6% -> 49.3% | Helped |
| Increase code length from 1024 to 4096 bits | FBE · Holidays+1M | mAP: 45.6% -> 47.2% | Helped |
| Increase code length from 4096 to 8192 bits | FBE · Holidays+1M | mAP: 47.2% -> 49.3% | Helped |
| Increase code length from 1024 to 4096 bits | FBE · CIFAR-10 | top-1 accuracy: 79.7% -> 80.7% | Helped |
| Increase code length from 4096 to 8192 bits | FBE · CIFAR-10 | top-1 accuracy: 80.7% -> 81.6% | Helped |
| Increase code length from 8192 to 16384 bits | FBE · CIFAR-10 | top-1 accuracy: 81.6% -> 82.4% | Helped |
| Use a fixed penalty weight rather than gradually increasing it (ablation) | FBE optimization · features across the reported datasets | accuracy | No clear effect |
| Use 16384-bit codes rather than 8192-bit codes | FBE · Holidays+1M | mAP | No clear effect |
Closest papers by meaning
- BiGR: Harnessing Binary Latent Codes for Image Generation and Improved Visual Representation Capabilities (ICLR 2025 · similarity 0.83)
- Image and Video Tokenization with Binary Spherical Quantization (ICLR 2025 · similarity 0.82)
- G-Net: A Provably Easy Construction of High-Accuracy Random Binary Neural Networks (NeurIPS 2025 · similarity 0.81)
- Binary Quadratic Quantization: Beyond First-Order Quantization for Real-Valued Matrix Compression (NeurIPS 2025 · similarity 0.81)
- Good, Cheap, and Fast: Overfitted Image Compression with Wasserstein Distortion (CVPR 2025 · similarity 0.81)
- Wavelet-Driven Masked Image Modeling: A Path to Efficient Visual Representation (AAAI 2025 · similarity 0.81)
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.