AI Paper Atlas

Research map / Architectures and efficiency / Quantization

On the Impact of Weight Quantization on Deep Neural Network Uncertainty

Shuang Liang, Xun Lu, Zi-Ang Liu, Ming-Liang Wang, Yan Lyu, Shao-Qun Zhang

AAAI 2026 · Architectures and efficiency › Quantization · approach: Mixed-precision quantization · 0 citations (OpenAlex)

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 changedWhereMeasuredEffect
apply MOMA to ternary MLP-2 trainingT-MLP-2 · MNISTuncertainty: 1.2572 -> 4.0129Improved
apply MOMA to ternary LeNet-5 trainingT-LeNet-5 · MNISTuncertainty: 2.8113 -> 4.0906Improved
apply MOMA to ternary VGG-13 trainingT-VGG-13 · MNISTuncertainty: 1.4500 -> 1.1463Improved
apply MOMA to ternary ResNet-18 trainingT-ResNet-18 · MNISTuncertainty: 4.2808 -> 2.4067Improved
apply MOMA to ternary VGG-13 trainingT-VGG-13 · MNISTaccuracy: 0.9954 -> 0.9957Improved
apply MOMA to ternary ResNet-18 trainingT-ResNet-18 · MNISTaccuracy: 0.9938 -> 0.9951Improved
apply MOMA to ternary MLP-2 trainingT-MLP-2 · MNISTaccuracy: 0.9752 -> 0.9751No clear effect
apply MOMA to ternary LeNet-5 trainingT-LeNet-5 · MNISTaccuracy: 0.9903 -> 0.9885Hurt
ternarize LeNet-5 weightsFP-LeNet-5 · MNISTuncertainty: 1.7728 -> 2.8113Hurt
ternarize ResNet-18 weightsFP-ResNet-18 · MNISTuncertainty: 6.9757 -> 4.2808Hurt
use EMP instead of MP for uncertainty estimationMP · MNIST (FP-LeNet-5)variance ratio vs MC: 1.1809 -> 1.0007Improved
use EMP instead of PL-DNN for uncertainty estimationPL-DNN · CIFAR-10 (FP-LeNet-5)variance ratio vs MC: 1.6167 -> 1.0098Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy