AI Paper Atlas

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

ARGENT: Adaptive Hierarchical Image-Text Representations

Chuong Huynh, Hossein Souri, Abhinav Kumar, Vitali Petsiuk, Deen Dayal Mohan, Suren Kumar

ECCV 2026 · Vision-language and multimodal › CLIP and zero-shot · approach: Recaptioning and text-supervised pretraining · 3 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
replace vanilla entailment loss with AdaEnt (ablation)MERU-S · CC3M ablationPEP AUC: 73.5 -> 95.5Improved
replace vanilla entailment loss with AdaEnt (ablation)MERU-S · CC3M ablationPEP AP: 45.0 -> 60.9Improved
replace vanilla entailment loss with AdaEnt (ablation)HyCoCLIP-S · CC3M ablationPEP AUC: 94.6 -> 97.3Improved
replace vanilla entailment loss with AdaEnt (ablation)HyCoCLIP-B · CC3M ablationPEP AUC: 94.2 -> 97.9Improved
replace vanilla entailment loss with AdaEnt (ablation)HyCoCLIP-L · CC3M ablationPEP AP: 76.0 -> 79.2Improved
replace vanilla entailment loss with AdaEnt (ablation)HyCoCLIP-S · CC3M ablationclassification accuracy: 22.9 -> 22.8No clear effect
apply adaptive weight h only to intra-modality pairs rather than all pairs (ablation)HyCoCLIP-S · CC3M ablationPEP AUC: 71.1 -> 97.3Improved
apply adaptive weight h to intra-modality pairs vs none (ablation)HyCoCLIP-S · CC3M ablationPEP AUC: 94.9 -> 97.3Improved
add regularization loss L_Reg (ablation)HyCoCLIP-S · CC3M ablationPEP AUC: 96.2 -> 97.3Improved
add regularization loss L_Reg (ablation)MERU-S · CC3M ablationclassification accuracy: 20.8 -> 20.6Mixed
replace HyCoCLIP entailment loss with AdaEnt (ARGENT)HyCoCLIP-S · ImageNet zero-shotaccuracy: 37.3 -> 38.8Improved
replace HyCoCLIP entailment loss with AdaEnt (ARGENT)HyCoCLIP-S · 16-dataset zero-shot avgaccuracy: 39.2 -> 38.4Hurt

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