AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Interpreting Deep Visual Representations via Network Dissection

Bolei Zhou, David Bau, Aude Oliva, Antonio B. Torralba

TPAMI 2018 · Label-efficient and robust learning › Recognition and long-tail · approach: Self-explaining and visual explanation networks · 348 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 random orthogonal rotation to conv5 representation (ablation)AlexNet Places205 conv5 · unique detectorsnumber of unique detectors: 80% fewerHurt
rotate pool5 representation then retrain fc layers (ablation)AlexNet Places365 · Places365 valtop-1 accuracy: 51.9 -> 51.7No clear effect
add batch normalization (ablation)Places205-AlexNet · conv5 interpretabilityunique detectorsHurt
add batch normalization (ablation)Places205-AlexNet · Places205 valtop-1 accuracy: 50.0 -> 50.5No clear effect
add dropout in FC layers (ablation)Places205-AlexNet · conv5 interpretabilityobject detectorsImproved
add dropout in FC layers (ablation)Places205-AlexNet · Places205 valtop-1 accuracy: 49.2 -> 50.0Improved
change random initialization (ablation)Places205-AlexNet · conv5 interpretabilityunique detectorsNo clear effect
triple conv5 units (256 to 768) (ablation)AlexNet-GAP · Places365 conv5unique detectors: 256 -> 768Improved
raise conv5 units to 1024 and 2048 (ablation)AlexNet-GAP-Wide · Places365 conv5unique detectors: 1024 and 2048No clear effect
use deeper architecture (ResNet over AlexNet)supervised CNNs · Places and ImageNet last conv layerunique detectorsImproved
train on Places instead of ImageNetsupervised CNNs · last conv layerunique detectorsImproved
use self-supervised instead of supervised trainingAlexNet · conv5object detectorsHurt

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