Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Sampling-Free Learning of Bayesian Quantized Neural Networks
Jiahao Su, Milan Cvitkovic, Furong Huang
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 BQN instead of E-QNN | MLP · MNIST | NLL ($10^{-3}$): 546.6$ \pm $157.9 -> 130.0$\pm$3.5 | Improved |
| Use BQN instead of E-QNN | MLP · KMNIST | NLL ($10^{-3}$): 2385.6$\pm$432.3 -> 457.7$\pm$13.8 | Improved |
| Use BQN instead of E-QNN | MLP · Fashion-MNIST | NLL ($10^{-3}$): 2529.4$\pm$276.7 -> 417.3$\pm$8.1 | Improved |
| Use BQN instead of E-QNN | CNN · MNIST | NLL ($10^{-3}$): 425.3$\pm$61.8 -> 41.8$\pm$1.6 | Improved |
| Use BQN instead of E-QNN | CNN · KMNIST | NLL ($10^{-3}$): 3755.7$\pm$465.1 -> 295.5$\pm$1.4 | Improved |
| Use BQN instead of E-QNN | CNN · Fashion-MNIST | NLL ($10^{-3}$): 1610.7$\pm$158.4 -> 209.5$\pm$2.8 | Improved |
| Use BQN instead of E-QNN | CNN · CIFAR10 | NLL ($10^{-3}$): 7989.7 $ \pm $ 600.2 -> 530.6 $ \pm $ 23.0 | Improved |
| Use BQN instead of E-QNN | CNN · Fashion-MNIST | % Err.: 3.02$\pm$0.37 -> 4.65$\pm$0.15 | Hurt |
| Use BQN instead of E-QNN | CNN · MNIST | % Err.: 0.85$\pm$0.13 -> 0.85$\pm$0.06 | No clear effect |
| Use sampling-free analytic inference rather than Monte Carlo sampling (ablation) | BQN · MNIST, KMNIST, Fashion-MNIST | accuracy and NLL | No clear effect |
| Increase the regularization coefficient λ (ablation) | BQN · CNN evaluation | gap between analytic inference and MC sampling | Hurt |
| Remove the prior regularization term by setting λ = 0 (ablation) | BQN · CNN evaluation | posterior uncertainty | Hurt |
Closest papers by meaning
- Encoding the Latent Posterior of Bayesian Neural Networks for Uncertainty Quantification (TPAMI 2023 · similarity 0.88)
- Vector Quantized Bayesian Neural Network Inference for Data Streams (AAAI 2021 · similarity 0.88)
- A Study of Bayesian Neural Network Surrogates for Bayesian Optimization (ICLR 2024 · similarity 0.86)
- Uncertainty Quantification with the Empirical Neural Tangent Kernel (NeurIPS 2025 · similarity 0.86)
- Do Bayesian Neural Networks Actually Behave Like Bayesian Models? (ICML 2025 · similarity 0.86)
- Make Me a BNN: A Simple Strategy for Estimating Bayesian Uncertainty from Pre-trained Models (CVPR 2024 · similarity 0.85)
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.
