AI Paper Atlas

Research map / Vision-language and multimodal / CLIP and zero-shot

Know When to Abstain: Optimal Selective Classification with Likelihood Ratios

Alvin Heng, Harold Soh

ICLR 2026 · Vision-language and multimodal › CLIP and zero-shot · approach: Feature calibration for unseen-class detection

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
estimate separate correct/wrong distributions (Delta-MDS vs MDS)MDS · ImageNet variants avg (1K), DFN CLIPAURC (x1e-2): 13.9 -> 7.81Improved
estimate separate correct/wrong distributions (Delta-KNN vs KNN)KNN · ImageNet variants avg (1K), DFN CLIPAURC (x1e-2): 13.1 -> 7.32Improved
average log-distance over top-k neighbors instead of k-th distance (ablation)Delta-KNN · ImageNet variants avg (1K), DFN CLIPAURC (x1e-2): 7.45 -> 7.32Improved
linearly combine Delta-KNN with RLogDelta-KNN · ImageNet variants avg (1K), DFN CLIPAURC (x1e-2): 7.32 -> 6.43Improved
linearly combine Delta-MDS with RLogDelta-MDS · ImageNet variants avg (1K), EVAAURC (x1e-2): 4.18 -> 3.86Improved
pair Delta-KNN with RLog instead of MSP (ablation)Delta-KNN combination · ImageNet variants avg (1K), DFN CLIPAURC (x1e-2): 7.18 -> 6.43Improved
combine two distance scores instead of distance plus logit (ablation)Delta-KNN · ImageNet variants avg (1K), DFN CLIPAURC (x1e-2): 6.43 -> 7.46Hurt
combine MSP with RLog (double-logit) (ablation)RLog · ImageNet variants avg (1K), DFN CLIPAURC (x1e-2): 7.39 -> 7.35No clear effect
use only 1% labeled samples for Delta-MDS statistics (ablation)Delta-MDS-RLog · ImageNet-1K, DFN CLIPAURC (x1e-2): 4.13 -> 10.5Hurt
use only 0.1% labeled samples for Delta-KNN (ablation)Delta-KNN-RLog · ImageNet-1K, DFN CLIPAURC (x1e-2): 3.98 -> 4.81No clear effect
use Delta-MDS-MSP instead of RLog combination on language taskDelta-MDS · Amazon Reviews in-distribution, DistilBERTAURC (x1e-2): 12.4 -> 11.9Improved
use Energy scoreEnergy · ImageNet variants avg (1K), DFN CLIPAURC (x1e-2): 11.5 -> 24.8Hurt

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