AI Paper Atlas

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

Improving Vision and Language Concepts Understanding with Multimodal Counterfactual Samples

Chengen Lai, Shengli Song, Sitong Yan, Guangneng Hu

ECCV 2024 · Vision-language and multimodal › CLIP and zero-shot · approach: Self-distillation for vision-language pretraining · 6 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
put counterfactual samples in the same mini-batch as factuals instead of random sampling (ablation)COMO · VL-Checklist averageaccuracy (%): 70.95 -> 76.93Improved
fine-tune on factual COCO only (ablation)CLIP · Winoground / VL-Checklistaverage score: 70.67 -> 70.84 (VL-Checklist avg)No clear effect
add counterfactual text to counterfactual image training (ablation)COMO · VL-Checklist averageaccuracy (%): 75.07 -> 76.93Improved
add counterfactual image to counterfactual text training (ablation)COMO · VL-Checklist averageaccuracy (%): 74.80 -> 76.93Improved
use weighted hard-negative contrastive loss instead of unweighted (ablation)COMO pretrained · VL-Checklist averageaccuracy (%): 76.08 -> 76.93Improved
use CLIP pretrained initialization instead of training from scratch (ablation)COMO · VL-Checklist averageaccuracy (%): 68.07 -> 76.93Improved
increase number of generated images per caption up to about 8 (ablation)COMO · VL-ChecklistaccuracyImproved
train COMO on 10% of data in low-resource setting (ablation)COMO · VL-Checklistrelative improvement over CLIP: 2.24% at 10% trainingImproved
COMO fine-tuning vs CLIP on 21 zero-shot classificationCOMO · 21 zero-shot classification datasetsaverage accuracy (%): 56.37 -> 55.87Hurt

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