AI Paper Atlas

Research map / Vision-language and multimodal / Cross-modal learning

Better Together: Leveraging Unpaired Multimodal Data for Stronger Unimodal Models

Sharut Gupta, Shobhita Sundaram, Chenyu Wang, Stefanie Jegelka, Phillip Isola

ICLR 2026 · Vision-language and multimodal › Cross-modal learning · approach: Multimodal fusion with missing modalities

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 unpaired text with shared weights (UML)self-supervised unimodal model · MUSTARDaccuracy: 59.66 -> 63.28Improved
add unpaired text with shared weights (UML)self-supervised unimodal model · UR-FUNNYaccuracy: 56.99 -> 57.34No clear effect
add unpaired text (UML)DINOv2 ViT-S/14 full fine-tuning · 9 vision datasets averageaccuracy: 81.54 -> 83.99Improved
add unpaired text (UML)DINOv2 ViT-S/14 full fine-tuning · SUN397accuracy: 66.20 -> 66.03Hurt
add unpaired text (UML)DINOv2 linear probe · 9 vision datasets average, 1-shotaccuracy: 45.52 -> 51.36Improved
add unpaired text (UML)DINOv2 linear probe · 9 vision datasets average, 4-shotaccuracy: 65.84 -> 68.53Improved
use CLIP aligned encoders instead of DINOv2 + OpenLLaMA (ablation)UML · image classificationaccuracyImproved
train with semantically unrelated modality (ablation)UML · image classificationaccuracyNo clear effect
unfreeze the text encoder (ablation)UML · image classificationaccuracyMixed
initialize ViT transformer with pretrained BERT weightsViT · ImageNetaccuracyImproved
add audio, vision and text auxiliary modalities (ablation)UML · ImageNet-ESC audio classificationaccuracyImproved

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