AI Paper Atlas

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

Improving Confidence Estimates for Unfamiliar Examples

Zhizhong Li, Derek Hoiem

CVPR 2020 · Label-efficient and robust learning › Recognition and long-tail · approach: Group-robust classification · 37 citations (OpenAlex)

Check 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
Apply temperature scaling calibrated on familiar datasingle model · Novel test splits across the evaluated image datasetsNLLImproved
Use an ensemble with temperature-scaled membersensemble · Novel test splitsNLLImproved
Distill ensemble predictions on familiar labeled and general unlabeled examplesG-distillation · Novel test splitsNLLImproved
Reduce predictions toward the familiar class prior according to novelty scoresNCR · Novel test splitsNLLMixed
Use G-distillation rather than standard distillationdistilled model · VOC-COCONLLNo clear effect
Use G-distillation rather than temperature scalingsingle model · VOC-COCO novel splitNLLImproved
Use G-distillation rather than temperature scalingsingle model · VOC-COCO familiar splitNLLHurt
Add an ensemble of G-distilled models (ablation)G-distillation · Novel test splitsNLLImproved
Add an ensemble of G-distilled models instead of using a calibrated ensemble (ablation)G-distillation · Novel test splitsNLLHurt

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