AI Paper Atlas

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

Toward Calibrated Mixture-of-Experts Under Distribution Shift

Gina Wong, Drew Prinster, Suchi Saria, Rama Chellappa, Anqi Liu

ICML 2026 · Label-efficient and robust learning › Recognition and long-tail · approach: Confidence calibration objectives · 0 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 entropy-balanced adversarial reweighting (Robust MoE)soft-routed MoE · CIFAR-10H hard subsetECE: 0.281 -> 0.074Improved
Add entropy-balanced adversarial reweighting (Robust MoE)soft-routed MoE · CivilComments hard subsetECE: 0.108 -> 0.037Improved
Add entropy-balanced reweightingsoft-routed MoE · CIFAR-10Haccuracy: 0.920 -> 0.904Hurt
Calibrate each expert post hoc (MoCaE)Vanilla MoE · CIFAR-10H hard subsetECE: 0.281 -> 0.262Improved
Calibrate each expert post hoc (MoCaE)Vanilla MoE · CIFAR-10H, CivilCommentsaccuracyNo clear effect
Apply robust term only to routing-relevant examples with ERM term kept (Robust Filtered vs Robust MoE) (ablation)Robust MoE · CIFAR-10HECE: 0.090 -> 0.013Improved
Apply robust term only to routing-relevant examples (Robust Filtered vs Robust MoE) (ablation)Robust MoE · CIFAR-10H hard subsetECE: 0.074 -> 0.122Hurt
Add temperature scaling after trainingRobust methods · CivilComments hard subsetECE: 0.037 -> 0.029Improved
Combine FGR with Robust MoE (ablation)Robust MoE · CivilComments hard subsetECE: 0.037 -> 0.065Hurt
Combine FGR with Robust MoE (ablation)Robust MoE · CIFAR-10H hard subsetECE: 0.074 -> 0.065Improved
Add FGR gradient rectificationVanilla MoE · CivilComments hard subsetECE: 0.108 -> 0.065Improved
Add FGR gradient rectificationVanilla MoE · CIFAR-10H hard subsetECE: 0.281 -> 0.262Improved

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