AI Paper Atlas

Research map / Architectures and efficiency / Pruning and efficient tuning

MEST: Accurate and Fast Memory-Economic Sparse Training Framework on the Edge

Geng Yuan, Xiaolong Ma, Wei Niu, Zhengang Li, Zhenglun Kong, Ning Liu, Yifan Gong, Zheng Zhan, Chaoyang He, Qing Jin, Siyue Wang, Minghai Qin and 4 more

NeurIPS 2021 · Architectures and efficiency › Pruning and efficient tuning · approach: Sparse training from scratch · 42 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
add elastic mutation (decaying mutation ratio) (ablation)MEST · CIFAR-100 98% sparsityaccuracy (%): 62.58 -> 64.59Improved
add elastic mutation (ablation)MEST · CIFAR-10 90% sparsityaccuracy (%): 92.12 -> 92.56Improved
add soft memory bound (EM&S) on top of EM (ablation)MEST+EM · CIFAR-100 98% sparsityaccuracy (%): 64.59 -> 67.16Improved
add soft memory bound (EM&S) on top of EM (ablation)MEST+EM · CIFAR-10 98% sparsityaccuracy (%): 89.22 -> 90.51Improved
add soft memory bound (ablation)MEST+EM · ImageNet ResNet-50 90% sparsityTop-1 accuracy (%): 73.63 -> 75.00Improved
add soft memory bound (ablation)MEST+EM · ImageNet ResNet-50 80% sparsityTop-1 accuracy (%): 75.75 -> 75.73No clear effect
add data-efficient two-phase training (DE) (ablation)MEST+EM&S · ImageNet ResNet-50 80% sparsityTop-1 accuracy (%): 75.73 -> 75.70No clear effect
add data-efficient training (DE) (ablation)MEST 1.7x+EM&S · ImageNet ResNet-50 90% sparsityTop-1 accuracy (%): 76.13 -> 76.08No clear effect
add DE for extra speedup (ablation)MEST · CIFAR-100 ResNet-32 90% sparsitytraining speedup: 10% to 15%Improved
use MEST+EM&S instead of RigL at similar FLOPsRigL · ImageNet 90% sparsity ResNet-50Top-1 accuracy (%): 72.0 -> 72.37Improved
use channel (structured) sparsity instead of unstructured/fine-grained (ablation)MEST · CIFAR-100accuracyHurt
use block or pattern sparsity instead of unstructured (ablation)MEST · CIFAR-100 ResNet-32training acceleration: up to 2.3xImproved

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