AI Paper Atlas

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

Supercharging Imbalanced Data Learning With Energy-based Contrastive Representation Transfer

Junya Chen, Zidi Xiu, Benjamin Goldstein, Ricardo Henao, Lawrence Carin, Chenyang Tao

NeurIPS 2021 · Label-efficient and robust learning › Recognition and long-tail · approach: Contrastive learning for imbalanced data · 8 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
Use multiple class-dependent source priors instead of one shared prior (ablation)ECRT · CIFAR100Top-1 accuracy: 52.31 -> 53.00Improved
Use multiple class-dependent source priors instead of one shared prior (ablation)ECRT · iNaturalistTop-1 accuracy: 68.38 -> 69.01Improved
Use multiple class-dependent source priors instead of one shared prior (ablation)ECRT · TinyImageNetTop-1 accuracy: 62.46 -> 64.40Improved
Use multiple class-dependent source priors instead of one shared prior (ablation)ECRT · ArXivaccuracy: 48.33 -> 48.33No clear effect
Train the feature encoder on majority and minority classes rather than only majority classes (ablation)ECRT · CIFAR100Top-1 accuracy: 53.00 -> 51.79Hurt
Add causal augmentation (ablation)ECRT · Toy modelaccuracyImproved
Increase augmentation strength when minority sample size is small (ablation)ECRT · Toy modelaccuracyHurt
Model directly in source space rather than feature space (ablation)ECRT · Toy modellearning speedImproved
Use parametric augmentation rather than nonparametric coordinate shuffling (ablation)ECRT · Low-sample augmentationMMD distance to ground truthImproved

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