AI Paper Atlas

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

Boosting Medical Visual Understanding From Multi-Granular Language Learning

Zihan Li, Yiqing Wang, Sina Farsiu, Paul Kinahan

ICLR 2026 · Vision-language and multimodal › CLIP and zero-shot · approach: CLIP training recipes and data curation · 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
Replace best baseline with MGLL pretrainingUniChest · MIDRC-XR (linear probe)AUC: 59.02 -> 61.25Improved
Pretrain with MGLL instead of CLIPCLIP · MIDRC-XR (full fine-tune)AUC: 88.52 -> 99.08Improved
Use MGLL vision encoder in LLaVA-MedLLaVA-Med · fundus multiple choiceaverage accuracy: 24.28 -> 58.37Improved
Use MGLL vision encoder in InternVLInternVL · fundus multiple choiceaverage accuracy: 77.35 -> 81.96Improved
Add point-wise loss to CLIP (ablation)CLIP · RFMiD linear probeAUC: 44.66 -> 70.34Improved
Add soft CLIP loss to CLIP (ablation)CLIP · RFMiD linear probeAUC: 44.66 -> 67.86Improved
Combine soft CLIP and point-wise losses (ablation)MGLL · RFMiD linear probeAUC: 70.34 -> 75.73Improved
Add smooth KL loss (ablation)MGLL · RFMiD linear probeAUC: 75.73 -> 79.62Improved
Use three granularity levels instead of one (ablation)MGLL · MIDRC-XR-Portable linear probeAUC: 80.54 -> 83.86Improved
Use ViT-L/14 instead of ViT-H/14 image encoder (ablation)MGLL · RFMiD linear probeAUC: 79.18 -> 79.62No clear effect
Use BERT instead of LLaMA text encoder (ablation)MGLL · RFMiD linear probeAUC: 74.89 -> 79.62Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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