AI Paper Atlas

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

Adaptive Conformal Guidance for Learning under Uncertainty

Rui Liu, Peng Gao, Yu Shen, Ming Lin, Pratap Tokekar

ICLR 2026 · Label-efficient and robust learning › Semi-supervised learning · approach: Pseudo-labeling for specialized tasks

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 AdaConG conformal weighting to KD with a noisy teacherKD, ResNet110 teacher to ResNet20 · CIFAR-100 with Gaussian noise domain shifttop-1 accuracy: 57.23 -> 66.53Improved
plain KD from an underperforming teacherResNet20 student · CIFAR-100 noisytop-1 accuracy: 66.51 (scratch) -> 57.23 (KD)Hurt
add AdaConG to feature-based KD (FitNet, PKT, FT)KD methods · CIFAR-100 noisy homogeneoustop-1 accuracy: FT ResNet110: +0.11 (below 0.15 threshold); others +0.26 to +2.41Improved
add AdaConG to semi-supervised methodsFixMatch · CIFAR-10 40 labelstop-1 accuracy: 64.18 -> 70.16Improved
add AdaConG to FlexMatchFlexMatch · CIFAR-100 2500 labelstop-1 accuracy: 63.59 -> 69.22Improved
use MSE instead of cross-entropy as guidance loss in SSL (ablation)AdaConG in UDA/FixMatch/FlexMatch · CIFAR-100top-1 accuracy: FixMatch 400 labels 8.56 -> 9.82No clear effect
use hard weighting (w=1 only when prediction set is a singleton) (ablation)AdaConG · CIFAR-100 KD ResNet110/ResNet20top-1 accuracy: 66.53 (soft) vs 66.52 (hard) for KDNo clear effect
use conformal quantile prediction sets instead of raw nonconformity score (ablation)AdaConG KD · CIFAR-100 ResNet110/ResNet20top-1 accuracy: 64.17 -> 66.53Improved
use conformal prediction uncertainty instead of heuristic MC dropout, MSP or entropyKD · CIFAR-100 ResNet110/ResNet20top-1 accuracy: 63.59 (MC dropout) -> 66.53Improved
increase temperature gamma in exponential weight up to 10 (ablation)AdaConG KD · CIFAR-100 ResNet110/ResNet20top-1 accuracyImproved
vary conformal error rate alpha (ablation)AdaConG · CIFAR-100 KD and CIFAR-10 FixMatchtop-1 accuracyNo clear effect
add AdaConG to KD for steering predictionKD · SullyChen steer prediction with noisemean accuracy: 73.5 -> 76.8Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 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