AI Paper Atlas

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

Learning What Matters: Prioritized Concept Learning via Relative Error-driven Sample Selection

Qian Yang, Shivam Chandhok, Oscar Mañas, Kanishk Jain, Aishwarya Agrawal, Leonid Sigal

CVPR 2026 · Vision-language and multimodal › CLIP and zero-shot · approach: Test-time adaptation of VLMs · 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
select clusters by relative accuracy improvement instead of easiest-first (ablation)PROGRESS · LLaVA-665K 20%relative performance (%): 92.3 -> 98.8Improved
select clusters by relative accuracy improvement instead of medium-first (ablation)PROGRESS · LLaVA-665K 20%relative performance (%): 91.1 -> 98.8Improved
select clusters by relative accuracy improvement instead of hardest-first (ablation)PROGRESS · LLaVA-665K 20%relative performance (%): 88.5 -> 98.8Improved
add progress-driven selection after warmup (ablation)PROGRESS · LLaVA-665K 20%relative performance (%): 94.6 -> 98.8Improved
use a random warmup instead of the default warmup (ablation)PROGRESS · LLaVA-665K 20%relative performance (%): 98.8 -> 97.1Mixed
lower softmax temperature from 1.0 to 0.3 (ablation)PROGRESS · LLaVA-665K 20%relative performance (%): 98.8 -> 92.8Hurt
raise softmax temperature to 1.2 (ablation)PROGRESS · LLaVA-665K 20%relative performance (%)Hurt
swap COINCIDE features for the unsupervised DINO-BERT features (ablation)COINCIDE · LLaVA-665K 20%relative performance (%): 97.8 -> 96.9No clear effect
use PROGRESS instead of COINCIDE selectionLLaVA-v1.5-7B · LLaVA-665K 20%relative performance (%): 97.8 -> 98.8Improved
apply PROGRESS to LLaVA-v1.5-13B without hyperparameter tuningLLaVA-v1.5-13B · LLaVA-665K 20%relative performance (%): 95.9 (COINCIDE) -> 96.8Improved
apply PROGRESS to Qwen2-VL-7BQwen2-VL-7B · LLaVA-665K 20%relative performance (%): 99.6 (COINCIDE) -> 100.0Improved
apply PROGRESS on the Vision-Flan datasetLLaVA-v1.5-7B · Vision-Flan-191K 16.7%relative performance (%): 95.8 (COINCIDE) -> 99.0Improved

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