AI Paper Atlas

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

KODA: Contrastive Representation Comparison and Alignment for Vision-Language Foundation Models

Youqi WU, Mohammad Jalali, Farzan Farnia

ICML 2026 · Vision-language and multimodal › Cross-modal learning · approach: Multimodal fusion with missing modalities · 0 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
fine-tune CLIP on KODA-selected slices with kernel alignment to DINOv2CLIP · FFHQ KODA slicesAMI: 0.25 -> 0.78Improved
fine-tune CLIP on KODA-selected slices with kernel alignment to BLIPCLIP · MSCOCO KODA slicesAMI: 0.52 -> 0.91Improved
fine-tune CLIP on KODA slices (ARI)CLIP · FFHQ KODA slicesARI: 0.19 -> 0.70Improved
evaluate on KODA-selected subset instead of full MSCOCOCLIP vs SigLIP retrieval · MSCOCO retrievalI2T R@1: CLIP 32.64 -> 18.00; SigLIP 42.64 -> 34.00Hurt
use joint image-text embeddings instead of image-only embeddings for KODA (ablation)KODA · MSCOCO discrepancy directionscluster coherence (qualitative)Improved
add explicit weak-clusterability constraint under reference kernel (vs kernel-difference of SPEC)KODA · contrastive embedding clusteringdiscrepancy discovery (qualitative)Improved

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