AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

Channel Regeneration: Improving Channel Utilization for Compact DNNs

Ankit Sharma, Hassan Foroosh

AAAI 2023 · Architectures and efficiency › RNNs and normalization · approach: Normalization layer theory · 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
Regenerate inhibited channels by resetting their BN gamma to the network median at regular intervalscompact CNNs · CIFAR-100accuracy: 69.10 -> 71.90Improved
Regenerate inhibited channels at regular intervalsMobileNetv2 · CIFAR-10accuracy: 90.74 -> 92.56Improved
Regenerate inhibited channels at regular intervalscompact architectures · Tiny-ImageNetaccuracy: 41.87 -> 46.40Improved
Regenerate inhibited channels at regular intervalsMobileNetv2 · ImageNetaccuracy: 71.70 -> 72.08Improved
Regenerate inhibited channels during segmentation trainingMobileNetv2 · ADE20Ksemantic segmentation performance: 69.69 -> 70.82Improved
Set regenerated BN gamma to the network median rather than the mean (ablation)MobileNetv2 · CIFAR-100accuracy: 71.50 -> 71.90Improved
Set regenerated BN gamma to the network median rather than 0.2 (ablation)MobileNetv2 · CIFAR-100accuracy: 71.25 -> 71.90Improved
Apply channel regeneration iteratively instead of only once (ablation)MobileNetv2 · CIFAR-100accuracyImproved
Vary the regeneration interval (ablation)ResNet-34 · CIFAR-10accuracy: 95.28 -> 95.69Mixed
Raise BN gamma for inhibited channels to revive channel gradient updatespruned models · CIFAR-100accuracy: 69.00 -> 69.42Improved
Raise BN gamma for inhibited channels in a pruned networkResNet-18 pruned with FPGM · ImageNetaccuracy: 66.11 -> 66.38Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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