Research map / Label-efficient and robust learning / Noisy labels
Training Binary Neural Networks through Learning with Noisy Supervision
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 |
|---|---|---|---|
| replace sign-based weight binarization with learned binary weight mapping (ablation) | ResNet-20 binary network · CIFAR-10 | accuracy: 85.26 -> 85.36 | Helped |
| add unbiased noisy supervision to learned binary weight mapping (ablation) | ResNet-20 binary network · CIFAR-10 | accuracy: 85.36 -> 85.56 | Helped |
| fine-tune using LNS rather than simple fine-tuning | binary ResNet-18 · ImageNet | Top-1 accuracy: 52.8% -> 53.1% | Helped |
| increase the noisy-supervision loss weight alpha from 0 to around 1 (ablation) | LNS · CIFAR-10 validation | accuracy: 85.35% -> 85.53% | Helped |
| set the assumed noise rate too high (rho=0.2) (ablation) | LNS · CIFAR-10 validation | accuracy | Hurt |
| use LNS instead of Dorefa-Net binarization | ResNet-20 binary network · CIFAR-10 | accuracy: 85.06 -> 85.78 | Helped |
| use LNS instead of the Bireal-Net + PReLU baseline | binary ResNet-18 · ImageNet | Top-1 accuracy: 59.0% -> 59.4% | Helped |
| use LNS instead of XNOR-Net | binary AlexNet · ImageNet | Top-1 accuracy: 44.2% -> 44.4% | Helped |
Closest papers by meaning
- Understanding Neural Network Binarization with Forward and Backward Proximal Quantizers (NeurIPS 2023 · similarity 0.88)
- AdaSTE: An Adaptive Straight-Through Estimator To Train Binary Neural Networks (CVPR 2022 · similarity 0.87)
- Training binary neural networks with real-to-binary convolutions (ICLR 2020 · similarity 0.87)
- Training Binary Neural Networks via Gaussian Variational Inference and Low-Rank Semidefinite Programming (NeurIPS 2024 · similarity 0.86)
- Taming Binarized Neural Networks and Mixed-Integer Programs (AAAI 2024 · similarity 0.86)
- BiPer: Binary Neural Networks using a Periodic Function (CVPR 2024 · 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.