AI Paper Atlas

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

Hallucination Improves the Performance of Unsupervised Visual Representation Learning

Jing Wu, Jennifer Hobbs, Naira Hovakimyan

ICCV 2023 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Empirical contrastive visual learning · 14 citations (OpenAlex)

Read the paper

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 Hallucinator to MoCoV2MoCoV2 (ResNet-18) · CIFAR-10 linear evaltop-1 accuracy: 87.21 -> 89.23Improved
add Hallucinator to SimSiamSimSiam (ResNet-18) · STL-10 linear evaltop-1 accuracy: 87.79 -> 88.16Improved
add Hallucinator to MoCoV1MoCoV1 (ResNet-50) · ImageNet-1K linear eval (100 epochs)top-1 accuracy: 57.27 -> 59.17Improved
add Hallucinator to SimCLRSimCLR (ResNet-50) · ImageNet-1K linear eval (100 epochs)top-1 accuracy: 61.23 -> 61.71Improved
add all modules (center cropping, extrapolation, hallucination) to MoCoV2 (ablation)MoCoV2 (ResNet-50) · ImageNet-200 linear evaltop-1 accuracy: 62.57 -> 63.58Improved
add center cropping alone (ablation)MoCoV2 (ResNet-50) · ImageNet-200 linear evaltop-1 accuracy: 62.57 -> 62.58No clear effect
add asymmetric extrapolation alone (ablation)MoCoV2 (ResNet-50) · ImageNet-200 linear evaltop-1 accuracy: 62.57 -> 62.82Improved
add non-linear hallucination layers (n=2 or 3) instead of identity n=0 (ablation)MoCoV2 (ResNet-50) · ImageNet-200 linear evaltop-1 accuracy: 62.85 -> 63.58Improved
reduce center-crop ratio p below 0.5 (ablation)MoCoV2 (ResNet-50) · ImageNet-200 linear evaltop-1 accuracy: 63.58 -> 63.21Hurt
use negative extrapolation range (-0.5, 0.0) (ablation)MoCoV2 (ResNet-50) · ImageNet-200 linear evaltop-1 accuracy: 63.58 -> 62.80Hurt
add Hallucinator to MoCoV2 pretraining for COCO transferMoCoV2 · COCO instance segmentationAP^mk: 34.2 -> 34.6Improved

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