AI Paper Atlas

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

Escaping Saddle Points for Effective Generalization on Class-Imbalanced Data

Harsh Rangwani, Sumukh K Aithal, Mayank Mishra, Venkatesh Babu Radhakrishnan

NeurIPS 2022 · Label-efficient and robust learning › Recognition and long-tail · approach: Group-robust classification · 9 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 SAM with high rhoCE + DRW · CIFAR-10 LTaccuracy: 75.5 -> 80.6Improved
add SAM with high rhoCE + DRW · CIFAR-100 LTaccuracy: 41.0 -> 44.6Improved
add SAMLDAM + DRW · CIFAR-10 LTaccuracy: 77.5 -> 81.9Improved
add SAMVS loss · CIFAR-10 LTaccuracy: 78.6 -> 82.4Improved
add SAMCE + DRW · CIFAR-10 step imbalanceaccuracy: 72.2 -> 79.3Improved
add SAMLDAM + DRW · ImageNet-LTaccuracy: 49.9 -> 53.1Improved
add SAMLDAM + DRW · iNaturalist 2018accuracy: 67.5 -> 70.1Improved
add SAM without re-weighting (ablation)CE · CIFAR-10 LTtail accuracy: 52.3 -> 51.7No clear effect
add SAM without re-weighting (ablation)CE · CIFAR-10 LToverall accuracy: 71.7 -> 73.1Improved
add PGD noise (ablation)CE + DRW · CIFAR-10 LTaccuracy: 75.5 -> 77.2Improved
add LPF-SGD (ablation)CE + DRW · CIFAR-10 LTaccuracy: 75.5 -> 78.5Improved
increase SAM rho from 0.05 to 0.8 (ablation)CE + DRW + SAM · CIFAR-10 LTtail accuracy: 63% -> 73%Improved

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