AI Paper Atlas

Research map / Architectures and efficiency / CNN design

Network of Experts for Large-Scale Image Categorization

Karim Ahmed, Mohammad Haris Baig, Lorenzo Torresani

ECCV 2016 · Architectures and efficiency › CNN design · approach: Filter design and trained-filter analysis · 100 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 K=10 experts instead of flat base modelAlexNet-C100 · CIFAR100top-1 accuracy: 54.0 -> 56.2Improved
increase number of experts from K=10 to K=50 (ablation)NofE fully-balanced · CIFAR100top-1 accuracy: 56.2 -> 55.33Hurt
use only K=2 experts (ablation)NofE fully-balanced · CIFAR100top-1 accuracy: 53.3 vs base 54.0No clear effect
use elasso soft balancing instead of fully-balanced (ablation)NofE K=10 · CIFAR100top-1 accuracy: 56.2 -> 55.59Hurt
learn specialties jointly with generalist instead of spectral clustering (ablation)NofE K=10 · CIFAR100top-1 accuracy: 53.2 -> 56.2Improved
learn specialties jointly instead of random balanced partition (ablation)NofE K=10 · CIFAR100top-1 accuracy: 53.7 -> 56.2Improved
initialise trunk from generalist instead of random weights (ablation)NofE AlexNet-C100 · CIFAR100top-1 accuracy: 49.53 -> 56.2Improved
initialise trunk from generalist instead of random weights (ablation)NofE ResNet56-C100 · CIFAR100top-1 accuracy: 73.95 -> 76.24Improved
train generalist from scratch instead of finetuning from pretrained base model (ablation)NofE AlexNet-C100 · CIFAR100top-1 accuracy: 55.5 -> 56.2Improved
convert base CNN into NofE (AlexNet-Quick)AlexNet-Quick-C100 · CIFAR100top-1 accuracy: 37.94 -> 45.58Improved
convert base CNN into NofE (VGG11)VGG11-C100 · CIFAR100top-1 accuracy: 68.48 -> 69.27Improved
match base model parameters and depth of NofE instead of using NofE (ResNet56) (ablation)ResNet56-C100 · CIFAR100top-1 accuracy: 73.88 -> 76.24Improved

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