AI Paper Atlas

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

Beyond Language Modeling: An Exploration of Multimodal Pretraining

Shengbang Tong, David Fan, John Nguyen, Ellis L Brown II, Gaoyue Zhou, Shengyi Qian, Boyang Zheng, Théophane Vallaeys, Rob Fergus, Naila Murray, Marjan Ghazvininejad, Mike Lewis and 8 more

ICML 2026 · Vision-language and multimodal › Human and embodied modeling · approach: Vision-language pretraining · 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
add a global shared expert (ablation)MoE 256 experts, 16 active · Notesperplexity: 27.392 -> 27.249Improved
use per-modality shared experts instead of a global one (ablation)MoE 256 experts · Notesperplexity: 27.249 -> 27.161Improved
use per-modality shared experts instead of none (ablation)MoE 256 experts · GenEvalscore: 0.360 -> 0.367Improved
add loss centeringSigLIP 2 unified model · DPGBenchDPG score: 0.570 -> 0.612Improved
add loss centeringSigLIP 2 unified model · DCLM valperplexity: 15.06 -> 15.36Hurt
add loss centeringSD-VAE unified model · DPGBenchDPG score: 0.470 -> 0.480Improved
replace shared FFN with modality-specific FFN (ablation)Transfusion baseline · DCLM valperplexity: 15.93 -> 15.13Improved
replace shared FFN with modality-specific FFN (ablation)Transfusion baseline · DPGBenchDPG score: 0.45 -> 0.47Improved
use SigLIP 2 (RAE) instead of VAE encoders (ablation)modality-specific FFN model · DPGBenchDPG score: 0.47 -> 0.57Improved
use MoE for parameter separation (vs dense, MoT) (ablation)SigLIP 2 model · DPGBenchDPG score: 0.63Improved
switch from v-pred to x-pred (ablation)MoE + SigLIP 2 · DPGBenchDPG score: 0.63 -> 0.65Improved
train jointly with language (multimodal vs T2I-only)dense model · DPGBenchDPG score: 0.598 -> 0.622Improved

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