AI Paper Atlas

Research map / Architectures and efficiency / Quantization

Bitwidth-Adaptive Quantization-Aware Neural Network Training: A Meta-Learning Approach

Jiseok Youn, Jaehun Song, Hyung-Sin Kim, Saewoong Bahk

ECCV 2022 · Architectures and efficiency › Quantization · approach: Mixed-precision quantization · 11 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
use MEBQAT bitwidth-task meta-training instead of AdaBitsMobileNet-v2 · CIFAR-10, bitwidth (2,2)accuracy: 58.98 -> 78.50Improved
use MEBQAT instead of dedicated per-bitwidth QATMobileNet-v2 · CIFAR-10, bitwidth (2,2)accuracy: 84.40 -> 78.50Hurt
use MEBQAT instead of Any-precision DNN8-layer CNN · SVHN, bitwidth (2,2)accuracy: 94.94 -> 97.25Improved
use MEBQAT instead of Any-precision DNN8-layer CNN · SVHN, bitwidth (4,4)accuracy: 96.19 -> 97.62Improved
use MEBQAT instead of Any-precision DNNPre-activation ResNet-20 · CIFAR-10, bitwidth (2,2)accuracy: 93.97 -> 92.52Hurt
use MEBQAT instead of dedicated QAT at full precisionMobileNet-v2 · CIFAR-10, bitwidth (FP,FP)accuracy: 93.00 -> 89.24Hurt
use MEBQAT-MAML instead of FOMAML5-layer CNN · Omniglot 20-way 1-shot, bitwidth (2,2)few-shot accuracy: 25.97 -> 89.57Improved
use MEBQAT-MAML instead of FOMAML+QAT5-layer CNN · Omniglot 20-way 1-shot, bitwidth (2,2)few-shot accuracy: 62.09 -> 89.57Improved
use MEBQAT-MAML instead of FOMAML+QAT5-layer CNN · MiniImageNet 5-way 5-shot, bitwidth (2,2)few-shot accuracy: 62.25 -> 61.65No clear effect
use MEBQAT-PN instead of PN4-layer CNN · Omniglot 20-way 1-shot, bitwidth (2,2)few-shot accuracy: 31.58 -> 94.87Improved
use MEBQAT-PN instead of PN+QAT4-layer CNN · MiniImageNet 5-way 1-shot, bitwidth (2,2)few-shot accuracy: 50.06 -> 47.66Hurt
use MEBQAT-PN instead of PN at 3-bit4-layer CNN · MiniImageNet 5-way 5-shot, bitwidth (3,3)few-shot accuracy: 46.74 -> 66.16Improved

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