Research map / Probabilistic and scientific modeling / Molecules and proteins
Sparse Deep Learning: A New Framework Immune to Local Traps and Miscalibration
Yan Sun, Wenjun Xiong, Faming Liang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 prior annealing instead of the multiple-run BNN_Evidence procedure | Sparse BNN · Synthetic nonlinear regression | MSPE: 2.439(0.132) -> 2.428(0.297) | No clear effect |
| Use sparse BNN prior annealing instead of Spinn with LASSO penalty | Sparse neural network · Synthetic nonlinear regression | MSPE: 4.488(0.350) -> 2.428(0.297) | Improved |
| Use sparse BNN prior annealing instead of a dense unregularized DNN | DNN · CIFAR-10, ResNet-20 | ECE: 0.02627(0.0005) -> 0.0175(0.0011) | Improved |
| Use sparse BNN prior annealing instead of DPF | ResNet-20 · CIFAR-10, pruning ratio around 10% | test accuracy: 92.17(0.21) -> 92.30(0.16) | Improved |
| Use BNN_anneal instead of DPF | ResNet-20 · CIFAR-10, pruning ratio around 10% | ECE: 0.0294(0.0009) -> 0.0175(0.0011) | Improved |
| Use BNN_anneal instead of DPF | ResNet-20 · CIFAR-10, pruning ratio around 20% | JS-Distance: 7.7329(0.1400) -> 6.4435(0.2029) | Hurt |
| Use BNN_anneal instead of DPF | ResNet-32 · CIFAR-10, pruning ratio around 5% | ECE: 0.0276(0.0019) -> 0.0149(0.0006) | Improved |
| Use sparse BNN prior annealing instead of a dense DNN | ResNet-20 · CIFAR-10 | test accuracy: 92.93(0.04) -> 92.30(0.16) | Hurt |
| Use prior-annealed sparse BNN to construct prediction intervals | Sparse DNN · Synthetic nonlinear regression, 95% prediction intervals | coverage rate: 94.72\%(0.61\%) | No clear effect |
Closest papers by meaning
- Random Sparse Lifts: Construction, Analysis and Convergence of finite sparse networks (ICLR 2024 · similarity 0.89)
- Sparsity in Deep Learning: Pruning and growth for efficient inference and training in neural networks (JMLR 2021 · similarity 0.88)
- Sparsity May Cry: Let Us Fail (Current) Sparse Neural Networks Together! (ICLR 2023 · similarity 0.88)
- A Single-Step, Sharpness-Aware Minimization is All You Need to Achieve Efficient and Accurate Sparse Training (NeurIPS 2024 · similarity 0.87)
- Global Minimizers of $\ell^p$-Regularized Objectives Yield the Sparsest ReLU Neural Networks (NeurIPS 2025 · similarity 0.87)
- The Iterative Optimal Brain Surgeon: Faster Sparse Recovery by Leveraging Second-Order Information (NeurIPS 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.
