AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

A Greedy Algorithm for Quantizing Neural Networks

Eric Lybrand, Rayan Saab

JMLR 2021 · Architectures and efficiency › RNNs and normalization · approach: Theory of depth and width · 13 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
quantize with GPFQ instead of memoryless scalar quantization (MSQ)CNN, 2-bit alphabet · CIFAR10 test (best C_alpha)top-1 accuracy: 0.2800 -> 0.8036Improved
quantize with GPFQ instead of MSQCNN, 4-bit alphabet · CIFAR10 testtop-1 accuracy: 0.8443 -> 0.8888Improved
quantize with GPFQ instead of MSQVGG16 fully connected layers · ILSVRC2012 top-1accuracy: 0.69485 -> 0.70075Improved
quantize with GPFQ instead of MSQ at large alphabet scalarVGG16 · ILSVRC2012 top-1accuracy: 0.53855 -> 0.68335Improved
increase alphabet scalar C_alpha from 2 to 6 for GPFQ (ablation)CNN 2-bit · CIFAR10 testtop-1 accuracy: 0.7522 -> 0.5365Hurt
increase bit budget from log2(3) to 4 bits for GPFQ (ablation)CNN · CIFAR10 testtop-1 accuracy: 0.7487 -> 0.8888Improved
increase bit budget from log2(3) to 4 bits for MSQ (ablation)CNN · CIFAR10 testtop-1 accuracy: 0.1464 -> 0.8443Improved
increase alphabet scalar C_alpha from 3 to 5 for MSQ (ablation)VGG16 · ILSVRC2012 top-1accuracy: 0.69485 -> 0.53855Hurt
increase alphabet scalar C_alpha from 3 to 5 for GPFQ (ablation)VGG16 · ILSVRC2012 top-1accuracy: 0.70075 -> 0.68335Hurt

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