pith. machine review for the scientific record. sign in

Solving token gradi- ent conflict in mixture-of-experts for large vision-language model

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it

fields

cs.CV 1

years

2026 1

verdicts

UNVERDICTED 1

representative citing papers

SMoES: Soft Modality-Guided Expert Specialization in MoE-VLMs

cs.CV · 2026-04-27 · unverdicted · novelty 6.0

SMoES improves MoE-VLM performance and efficiency via soft modality-guided expert routing and inter-bin mutual information regularization, yielding 0.9-4.2% task gains and 56% communication reduction.

citing papers explorer

Showing 1 of 1 citing paper.

  • SMoES: Soft Modality-Guided Expert Specialization in MoE-VLMs cs.CV · 2026-04-27 · unverdicted · none · ref 63

    SMoES improves MoE-VLM performance and efficiency via soft modality-guided expert routing and inter-bin mutual information regularization, yielding 0.9-4.2% task gains and 56% communication reduction.