AI Paper Atlas

Research map / Architectures and efficiency / CNN design

Deep Feature Surgery: Towards Accurate and Efficient Multi-Exit Networks

Cheng Gong, Yao Chen, Qiuyang Luo, Ye Lu, Tao Li, Yuzhi Zhang, Yufei Sun, Le Zhang

ECCV 2024 · Architectures and efficiency › CNN design · approach: Efficient CNN scaling and dynamic inference · 10 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
apply Deep Feature Surgery to multi-exit ResNet18ResNet18 multi-exit · CIFAR-100 Exit1top-1 accuracy: 67.85 -> 74.17Improved
apply Deep Feature Surgery to multi-exit ResNet18 ensembleResNet18 multi-exit · CIFAR-100 Ensembletop-1 accuracy: 79.67 -> 80.89Improved
apply Deep Feature Surgery to MSDNetMSDNet · CIFAR-100 all exitstop-1 accuracy: 72.46 vs 68.78 (Exit1); 79.56 vs 77.47 (Exit7)Improved
apply Deep Feature Surgery to MSDNet on ImageNetMSDNet · ImageNet average of 5 exitstop-1 accuracy: 67.58 -> 69.05Improved
DFS versus Mate-GF at the deepest ImageNet exitMSDNet · ImageNet Exit5top-1 accuracy: 73.27 vs 72.88Hurt
lower partition coefficient beta (more exit-specific features) (ablation)DFS on VGG7-64 · CIFAR-100 early exitstop-1 accuracy: 17.88 -> 22.81 (Exit1, beta 0.9 -> 0.1)Improved
higher partition coefficient beta (more shared features) (ablation)DFS on VGG7-64 · CIFAR-100 late exitstop-1 accuracy: 65.46 -> 73.06 (Exit6, beta 0.1 -> 0.9)Improved
feature partitioning reduces backward operations in trainingDFS · VGG7-64 CPU trainingtraining FPS: up to +14.43% vs DSN, +50.00% vs BYOTImproved
DFS training on GPU with VGG16DFS · VGG16 GPU trainingtraining FPS: up to +14.99% vs DSN, +22.23% vs BYOTImproved
DFS in budgeted batch classification versus Meta-GFMSDNet · CIFAR-100 budgeted batchFLOPs for same accuracy: about 2x fewer FLOPsImproved

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