AI Paper Atlas

Research map / Architectures and efficiency / Mixture of experts

Soft Modality-Guided Expert Specialization in MoE-VLMs

Zi-Hao Bo, Yaqian Li, Anzhou Hou, Rinyoichi Takezoe, Ertao Zhao, Tianxiang Pan, Jiale Yan, Mo Guang, Kaiwen Long

CVPR 2026 · Architectures and efficiency › Mixture of experts · approach: Expert routing and specialization

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 hard 0/1 modality score with soft score (attention) (ablation)SMoES · DeepSeekMoE VLMoverall relative avg: -0.3% -> +3.5%Improved
add inter-bin MI loss on top of expert binning (ablation)SMoES · DeepSeekMoE VLMoverall relative avg: +1.7% -> +3.5%Improved
replace MI with KL divergence for inter-bin specialization (ablation)SMoES · DeepSeekMoE VLMoverall relative avg: +3.5% -> -4.1%Hurt
add expert binning alone (ablation)soft routing · DeepSeekMoE VLMoverall relative avg: +1.7%Improved
add expert binning alone (ablation)soft routing · OLMoE VLMoverall relative avg: -11.2%Hurt
use adaptive instead of fixed binning (ablation)SMoES attention-soft · DeepSeekMoE VLMoverall relative avg: +1.3% -> +3.5%Improved
add hard routing of experts by modalitysoft routing · DeepSeekMoE VLMoverall relative avg: -12.3% (t32-v32)Hurt
add hybrid shared/modality experts (MoIIE)soft routing · DeepSeekMoE VLMoverall relative avg: -4.8% to -9.3%Hurt
add SMAR KL regularizationsoft routing · DeepSeekMoE VLMoverall relative avg: -3.9% to -10.2%Hurt
use SMoES on Qwen3-MoE (gaussian-soft)soft routing · Qwen3-30B-A3B VLMlanguage-only avg: -0.2%No clear effect
use GMM with k=2 instead of single Gaussian (ablation)SMoES gaussian-soft · OLMoE VLMoverall relative avg: +2.0% -> +2.8%Improved
vary Gaussian temperature (0.05D, 1.0D vs 0.5D) (ablation)SMoES gaussian-soft · DeepSeekMoE VLMoverall relative avg: +2.4% -> +1.2% (0.05D), +1.1% (1.0D)Hurt

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