AI Paper Atlas

Research map / Label-efficient and robust learning / Semi-supervised learning

CaliMatch: Adaptive Calibration for Improving Safe Semi-supervised Learning

Jinsoo Bae, Seoung Bum Kim, Hyungrok Do

ICCV 2025 · Label-efficient and robust learning › Semi-supervised learning · approach: Confidence-based pseudo-label selection · 2 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 multiclass calibration (adaptive label smoothing + temperature scaling) (ablation)CaliMatch · CIFAR-100 (kappa 60%)accuracy (%): 66.16 -> 68.56Improved
add OOD detector calibration (ablation)CaliMatch · CIFAR-10 (kappa 60%)accuracy (%): 87.57 -> 87.62No clear effect
add OOD detector calibration (ablation)CaliMatch · CIFAR-100 (kappa 60%)accuracy (%): 66.96 -> 68.56Improved
add OOD detector calibration (ablation)CaliMatch · TinyImageNet unseen detection ECEECE: 0.238 vs 0.189 (w/o multiclass calibration)Mixed
add both calibrations (CaliMatch) vs OpenMatch (ablation)OpenMatch · CIFAR-100 (kappa 60%)accuracy (%): 65.65 -> 68.56Improved
add calibration to OpenMatch on ImageNetOpenMatch · ImageNet (kappa 60%)top-1 accuracy (%): +1.24% over OpenMatchImproved
add calibration to OpenMatch on ImageNetOpenMatch · ImageNet (kappa 60%)OOD ECE: 0.121 -> 0.070Improved
raise OOD threshold tau_1 from 0.7 to 0.8 (ablation)CaliMatch · CIFAR-10accuracyHurt
use CaliMatch vs best baseline on CIFAR-10 kappa 30%CaliMatch · CIFAR-10 (kappa 30%)top-1 accuracy (%): 90.30 (SCOMatch) vs 90.25No clear effect
vary lambda_O and lambda_OCal (ablation)CaliMatch · CIFAR-10 (kappa 60%)accuracyNo clear effect

Later papers that compared with this method

2 later accepted papers report CaliMatch as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 4 on most shared results.

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