Research map / Optimization / Optimizers
AdaSTE: An Adaptive Straight-Through Estimator To Train Binary Neural Networks
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 AdaSTE instead of MD-tanh | binary-weight VGG-16 · CIFAR-10 | accuracy: 91.64 -> 92.37 | Helped |
| use AdaSTE instead of MD-softmax-s | binary-weight ResNet-18 · CIFAR-10 | accuracy: 93.28 -> 94.11 | Helped |
| use AdaSTE instead of PMF | binary-weight VGG-16 · CIFAR-100 | accuracy: 64.71 -> 69.28 | Helped |
| use AdaSTE instead of MD-softmax-s | binary-weight ResNet-18 · CIFAR-100 | accuracy: 72.18 -> 75.03 | Helped |
| use AdaSTE instead of BayesBiNN | binary-weight ResNet-18 · TinyImageNet | accuracy: 54.22 -> 54.92 | Helped |
| add annealing to AdaSTE (ablation) | binary-weight VGG-16 · CIFAR-10 | accuracy: {92.16}$\pm$0.16 -> 92.37$\pm$0.09 | Helped |
| add annealing to AdaSTE (ablation) | binary-weight ResNet-18 · CIFAR-10 | accuracy: 93.96$\pm$0.14 -> 94.11$\pm$0.08 | Helped |
| add annealing to AdaSTE (ablation) | binary-weight VGG-16 · CIFAR-100 | accuracy: 68.46$\pm$0.18 -> 69.28$\pm$0.17 | Helped |
| add annealing to AdaSTE (ablation) | binary-weight ResNet-18 · CIFAR-100 | accuracy: 73.90$\pm$0.20 -> 75.03$\pm$0.35 | Helped |
| add annealing to AdaSTE (ablation) | binary-weight ResNet-18 · TinyImageNet | accuracy: 53.49 -> 54.92 | Helped |
| apply mixup (ablation) | AdaSTE · TinyImageNet, ResNet-18 | accuracy: 54.92 -> 56.11 | Helped |
| apply mixup (ablation) | AdaSTE · Imagenette, ResNet-18 | accuracy: 79.66 -> 80.91 | Helped |
Closest papers by meaning
- An Empirical study of Binary Neural Networks' Optimisation (ICLR 2019 · similarity 0.91)
- Training Binary Neural Networks through Learning with Noisy Supervision (ICML 2020 · similarity 0.87)
- BEP: A Binary Error Propagation Algorithm for Binary Neural Networks Training (ICLR 2026 · similarity 0.87)
- Training Binary Neural Networks via Gaussian Variational Inference and Low-Rank Semidefinite Programming (NeurIPS 2024 · similarity 0.87)
- Layerwise Progressive Freezing: A Training Scaffold for Depth-Scalable Binary Networks (NeurIPS 2026 · similarity 0.85)
- Fast and Slow Gradient Approximation for Binary Neural Network Optimization (AAAI 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.