Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Masked Bayesian Neural Networks : Theoretical Guarantee and its Posterior Inference
Insung Kong, Dongyoon Yang, Jongjin Lee, Ilsang Ohn, GYUSEUNG BAEK, Yongdai Kim
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 learned node masks rather than a dense BNN | mBNN · Boston, Concrete, Energy, Yacht | RMSE: Boston: 3.411(0.145) -> 2.902(0.143); Concrete: 5.080(0.178) -> 4.913(0.180); Energy: 0.591(0.017) -> 0.474(0.015); Yach | Improved |
| Use learned node masks rather than a dense BNN | mBNN · Boston, Concrete, Energy, Yacht | NLL: Boston: 2.726(0.087) -> 2.472(0.085); Concrete: 3.091(0.055) -> 3.027(0.046); Energy: 0.902(0.025) -> 0.670(0.034); Yach | Improved |
| Use learned node masks rather than a dense BNN | mBNN · Boston, Concrete, Energy, Yacht | CRPS: Boston: 1.738(0.052) -> 1.462(0.055); Concrete: 2.658(0.072) -> 2.628(0.087); Energy: 0.322(0.007) -> 0.256(0.006); Yach | Improved |
| Use masked Bayesian convolution rather than node-sparse deep ensembles | mBCNN · CIFAR10 and CIFAR100 | ACC: CIFAR10: 0.926(0.003) -> 0.932(0.001); CIFAR100: 0.735(0.003) -> 0.737(0.001) | Mixed |
| Use masked Bayesian convolution rather than node-sparse deep ensembles | mBCNN · CIFAR10 and CIFAR100 | NLL: CIFAR10: 0.255(0.011) -> 0.220(0.005); CIFAR100: 1.076(0.014) -> 1.004(0.008) | Improved |
| Use masked Bayesian convolution rather than node-sparse deep ensembles | mBCNN · CIFAR10 and CIFAR100 | FLOPs: CIFAR10: 22.45(1.64)% -> 12.26(0.06)%; CIFAR100: 27.02(2.30)% -> 18.34(0.27)% | Improved |
| Use less-sensitive-node death proposals instead of linear-approximation proposals (ablation) | mBNN MH sampler · MCMC burn-in; activated-node ratio | activated-node reduction speed | Improved |
| Use a linear approximation of the log-posterior for mask proposals (ablation) | mBNN MH sampler · MCMC burn-in; activated-node ratio | activated-node reduction speed | Hurt |
Closest papers by meaning
- Posterior Concentrations of Fully-Connected Bayesian Neural Networks with General Priors on the Weights (JMLR 2025 · similarity 0.92)
- Bayesian Optimization for Sparse Neural Networks With Trainable Activation Functions (TPAMI 2024 · similarity 0.89)
- A Study of Bayesian Neural Network Surrogates for Bayesian Optimization (ICLR 2024 · similarity 0.88)
- Training Bayesian Neural Networks with Sparse Subspace Variational Inference (ICLR 2024 · similarity 0.88)
- Posterior Contraction for Sparse Neural Networks in Besov Spaces with Intrinsic Dimensionality (NeurIPS 2025 · similarity 0.87)
- Posterior Contraction Rates for sparse Kolmogorov-Arnold Networks in Anisotropic Besov Spaces (NeurIPS 2026 · 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.
