AI Paper Atlas

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

Max-Margin Contrastive Learning

Anshul Shah, Suvrit Sra, Rama Chellappa, Anoop Cherian

AAAI 2022 · Label-efficient and robust learning › Recognition and long-tail · approach: Contrastive representation theory · 42 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
replace the contrastive loss with MMCLImageNet-100 linear evaluation · ImageNet-100top-1: 79.0 -> 80.7Improved
replace the contrastive loss with MMCLImageNet-1K linear evaluation · ImageNet-1Ktop-1: 63.9 -> 63.8No clear effect
replace SimCLR loss with MMCLSTL-10 with batch size 128 · STL-10accuracy: 77.18 -> 86.8Improved
replace SimCLR loss with MMCLCIFAR-100 with batch size 512 · CIFAR-100accuracy: 66.32 -> 70.7Improved
replace MoCo-v2 with MMCLUCF-101 RGB · UCF-101top-1: 46.8 -> 52.45Improved
replace GraphCL loss with MMCLMUTAG graph representation learning · MUTAGaccuracy: 86.80 ± 1.34 -> 88.42 ± 1.33Improved
replace the RBF kernel with a linear kernel (ablation)MMCL · CIFAR-100readout accuracy: 55.35 % -> 41.43%Hurt
replace the RBF kernel with a tanh kernel (ablation)MMCL · CIFAR-100readout accuracy: 55.35 % -> 54.53%Hurt
replace the RBF kernel with a linear kernel (ablation)MMCL · STL-10readout accuracy: 81.33% -> 74.82%Hurt
replace the RBF kernel with a tanh kernel (ablation)MMCL · STL-10readout accuracy: 81.33% -> 80.5%Hurt
use a small slack penalty C=0.1 (ablation)MMCL trained on CIFAR-100 · CIFAR-100readout accuracy: 55.1 -> 44.41Hurt
use a small slack penalty C=0.1 (ablation)MMCL trained on STL-10 · STL-10readout accuracy: 79.42 -> 73.82Hurt

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