AI Paper Atlas

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

Unsupervised Learning under Latent Label Shift

Manley Roberts, Pranav Mani, Saurabh Garg, Zachary Chase Lipton

NeurIPS 2022 · Label-efficient and robust learning › Recognition and long-tail · approach: Learning from unlabeled and positive-unlabeled data · 2 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 domain information via DDFA (SI) instead of SCAN aloneDDFA (SI) · CIFAR-20 r=20, alpha 0.5, kappa 8test accuracy: 0.445 -> 0.807Improved
Use DDFA (SI) instead of SCAN in the densest settingDDFA (SI) · CIFAR-20 r=20, alpha 10, kappa 20test accuracy: 0.438 -> 0.355Hurt
Initialize DDFA backbone with SCAN weights (SI) instead of random (RI) (ablation)DDFA · CIFAR-20 r=20, alpha 0.5, kappa 8test accuracy: 0.549 -> 0.807Improved
Initialize DDFA backbone with SCAN weights (SI) instead of random (RI) (ablation)DDFA · CIFAR-20 r=20, alpha 10, kappa 20test accuracy: 0.203 -> 0.355Improved
Make label marginals denser (alpha 0.5 to alpha 10)DDFA (SI) · CIFAR-20 r=20test accuracy: 0.807 -> 0.355Hurt
Increase number of domains r from 20 to 30DDFA (SI) · CIFAR-20, alpha 0.5, kappa 8test accuracy: 0.807 -> 0.868Improved
Increase number of domains r from 20 to 30DDFA (SI) · CIFAR-20, alpha 10, kappa 20test accuracy: 0.355 -> 0.517Improved
Add domain information via DDFA (SPI) instead of SCANDDFA (SPI) · FieldGuide-2 r=2, alpha 0.5, kappa 3test accuracy: 0.589 -> 0.947Improved
Use DDFA (SPI) instead of SCAN in the densest settingDDFA (SPI) · FieldGuide-2 r=2, alpha 10, kappa 7test accuracy: 0.601 -> 0.676Improved
Use DDFA (SPI) instead of SCAN on label marginal errorDDFA (SPI) · FieldGuide-2 r=2, alpha 0.5, kappa 3Q_Y|D err: 0.880 -> 0.084Improved
Use DDFA (SPI) instead of SCAN on label marginal error with r=5DDFA (SPI) · FieldGuide-2 r=5, alpha 10, kappa 7Q_Y|D err: 0.226 -> 0.253Hurt
Use DDFA without SCAN on easy datasetDDFA · CIFAR-10test accuracyHurt

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