AI Paper Atlas

Research map / Label-efficient and robust learning / Noisy labels

Learning from Label Proportions by Learning with Label Noise

Jianxin Zhang, Yutong Wang, Clayton Scott

NeurIPS 2022 · Label-efficient and robust learning › Noisy labels · approach: Instance-dependent noise modeling · 3 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 KL loss with LLPFC-uniformWide ResNet-16-4 · CIFAR10, bag size 256test accuracy: .3749 $\pm$ .14 -> .7381 $\pm$ .01Improved
Replace KL loss with LLPFC-approxWide ResNet-16-4 · EMNIST, bag size 512test accuracy: .0845 $\pm$ .01 -> .9048 $\pm$ .00Improved
Replace KL loss with LLPFC-approxResNet18 · CIFAR10, bag size 32test accuracy: .7837 $\pm$ .01-> .6567 $\pm$ .01Hurt
Replace KL loss with LLPFC-approxResNet18 · CIFAR10, bag size 1024test accuracy: .5197 $\pm$ .04 -> .6169 $\pm$ .01Improved
Replace KL loss with LLPFC-approxResNet18 · EMNIST, bag size 32test accuracy: .9319 $\pm$ .00 -> .9146 $\pm$ .00Hurt
Replace KL loss with LLPFC-approxResNet18 · SVHN, bag size 32test accuracy: .1279 $\pm$ .06 -> .8824 $\pm$ .01Improved
Replace KL loss with LLPFC-uniformVGG16 · CIFAR10, bag size 32test accuracy: .2513 $\pm$ .11 -> .7602 $\pm$ .00Improved
Replace KL loss with LLPFC-uniformVGG16 · SVHN, bag size 256test accuracy: .1024 $\pm$ .05 -> .1734 $\pm$ .11No clear effect
Replace LLPGAN with LLPFC-approxLLPGAN discriminator · CIFAR10, bag size 32test accuracy: .3630 $\pm$ .01 -> .6169 $\pm$ .01Improved
Replace LLPGAN with LLPFC-uniformLLPGAN discriminator · SVHN, bag size 32test accuracy: .2378 $\pm$ .24 -> .8800 $\pm$ .00Improved
Use LLPFC-approx instead of LLPFC-uniformWide ResNet-16-4 · EMNIST, bag size 2048test accuracy: .8817 $\pm$ .01 -> .9007 $\pm$ .01Improved
Use LLPFC-approx instead of LLPFC-uniformVGG16 · SVHN, bag size 128test accuracy: .5402 $\pm$ .21 -> .1091 $\pm$ .04Hurt

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