AI Paper Atlas

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

AdaFocal: Calibration-aware Adaptive Focal Loss

Arindam Ghosh, Thomas Schaaf, Matthew R. Gormley

NeurIPS 2022 · Label-efficient and robust learning › Recognition and long-tail · approach: Confidence calibration objectives · 11 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 AdaFocal adaptive per-bin gamma instead of FLSD-53 heuristic gammafocal loss calibration · CIFAR-10 ResNet50ECE_EM (%): 1.63 -> 0.66Improved
use AdaFocal instead of FLSD-53 on a dataset where FLSD-53 is under-confidentfocal loss calibration · ImageNet ResNet50ECE_EM (%): 16.77 -> 1.87Improved
use AdaFocal instead of cross entropyclassifier training · CIFAR-100 ResNet50ECE_EM (%): 17.17 -> 1.36Improved
use AdaFocal instead of cross entropyclassifier training · CIFAR-10 ResNet50test error (%): 4.95 -> 5.30Hurt
use AdaFocal instead of cross entropyclassifier training · CIFAR-100 ResNet50test error (%): 23.30 -> 22.60Improved
use AdaFocal instead of cross entropyclassifier training · 20Newsgroup CNNtest error (%): 26.68 -> 28.53Hurt
add temperature scaling after AdaFocal trainingAdaFocal · CIFAR-10 ResNet50ECE_EM (%): 0.66 -> 0.44Improved
use 15 rather than 5 equal-mass bins for gamma statistics (ablation)AdaFocal · CIFAR-10 ResNet50ECE_EM (%): 0.63 -> 0.56Mixed
use 50 rather than 15 bins for gamma statistics (ablation)AdaFocal · CIFAR-10 ResNet50ECE_EM (%): 0.56 -> 1.10Hurt
update gamma more frequently (every few mini-batches instead of per epoch) (ablation)AdaFocal · 20Newsgroup CNN/BERTcalibration over trainingImproved
switch to inverse-focal loss when gamma falls below threshold (ablation)AdaFocal · ImageNet ResNet50ECE_EM (%)Improved

Later papers that compared with this method

1 later accepted paper reports AdaFocal as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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