AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

ModernVBERT: Towards Smaller Visual Document Retrievers

Paul Teiletche, Quentin Macé, Max Conti, António Loison, Gautier Viaud, Pierre Colombo, Manuel Faysse

ICML 2026 · Vision-language and multimodal › Human and embodied modeling · approach: Layout-aware document understanding · 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
use bidirectional attention (MLM encoder) instead of causal decoder for late-interaction retrieval (ablation)ModernVBERT ablation models · ViDoRe multi-vectornDCG@5: 54.2 -> 64.8 (+10.6)Improved
use bidirectional attention instead of causal attention for single-vector retrieval (ablation)ModernVBERT ablation models · ViDoRe single-vectornDCG@5No clear effect
remove causal mask / switch to MLM only during later training (dec-to-enc, dec with MLM) (ablation)decoder-based retriever · ViDoRe multi-vectornDCG@5No clear effect
early fusion of vision encoder with pretrained language model vs fine-tuned SigLIP2 dual encoder (ablation)vision-language retriever · document retrieval vs natural image tasksnDCG@5 / accuracy: +6.1 nDCG@5 on document retrieval; worse on natural image tasksMixed
scale modality alignment data (ablation)enc · document retrieval vs natural imagesretrieval scoreMixed
increase input image resolution from 512px to 2048px (ablation)enc · document retrievalnDCG@5: 30.7 -> 43.8Improved
increase input image resolution from 512px to 2048px (ablation)enc · image classificationaccuracy: 41.4 -> 33.9Hurt
add high-resolution cooldown at end of modality alignment (ablation)enc 2048px · document retrievalnDCG@5: 43.8 -> 45.8Improved
add text-only query-document pairs to contrastive mix (ablation)enc · document retrievalnDCG@5: 43.9 -> 45.6Improved
add text-only query-document pairs to contrastive mix (ablation)enc · image/caption retrievalscore: 57.2 -> 53.2Hurt
add NatCap synthetic image-caption pairs (ablation)enc · image classificationaccuracy: 35.7 -> 49.9Improved
add NatCap synthetic image-caption pairs (ablation)enc · document retrievalnDCG@5: 45.6 -> 45.8Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 0 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