AI Paper Atlas

Research map / Label-efficient and robust learning / Self-supervised and contrastive

Discriminability-Transferability Trade-Off: An Information-Theoretic Perspective

Quan Cui, Bingchen Zhao, Zhao-Min Chen, Borui Zhao, Renjie Song, Boyan Zhou, Jiajun Liang, Osamu Yoshie

ECCV 2022 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Empirical contrastive visual learning · 19 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 CTC two-stage contrastive temporal coding (ablation)ResNet-18 CosLr · CIFAR-100top-1 accuracy: 79.3 -> 80.1Improved
add CTCResNet-50 CosLr · ImageNettop-1 accuracy: 76.1 -> 76.4Improved
add CTC on top of AutoAugment (ablation)ResNet-18 · CIFAR-100top-1 accuracy: 80.4 -> 81.2Improved
add CTC on top of AutoAugment (ablation)ResNet-50 · ImageNettop-1 accuracy: 76.8 -> 77.2Improved
add AutoAugment to baseline training (ablation)ResNet-18 CosLr · CIFAR-100top-1 accuracy: 79.3 -> 80.4Improved
transfer CTC representations (frozen backbone)ResNet-50 · CUB200top-1 accuracy: 62.5 -> 63.7Improved
transfer AA+CTC representations (unfrozen backbone)ResNet-50 · FGVC-Aircrafttop-1 accuracy: 83.4 -> 85.6Improved
use CTC pre-training instead of CosLrResNet-50 · COCO detectionAP bbox: 38.2 -> 39.5Improved
use CTC instead of InfoMin self-supervised pre-trainingResNet-50 · COCO detectionAP mask 75 (AP bbox 75): 42.0 -> 42.0No clear effect
use CTC instead of MoCo v1 pre-trained on Instagram-1BResNet-50 · iNaturalist-18 (frozen)top-1 accuracy: 65.8 -> 66.4Improved
raise IAS loss weight alpha from 0.1 to 0.5 (ablation)CTC · CIFAR-100top-1 accuracyMixed

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