ExPLoRe turns MoE dispatch weights into per-patch loss coefficients for multi-objective masked image modeling, reporting gains on ImageNet-1K and ADE20K transfer.
arXiv preprint arXiv:2211.09799 , year=
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CV 2years
2026 2verdicts
UNVERDICTED 2representative citing papers
MACCO applies cross-modal masked reconstruction of compositional concepts with inter- and intra-modal auxiliary objectives to improve visio-linguistic compositionality in VLMs.
citing papers explorer
-
ExPLoRe: Expert Patch-Level Loss Routing for Multi-Objective Masked Image Modeling
ExPLoRe turns MoE dispatch weights into per-patch loss coefficients for multi-objective masked image modeling, reporting gains on ImageNet-1K and ADE20K transfer.
-
Cross-Modal Masked Compositional Concept Modeling for Enhancing Visio-Linguistic Compositionality
MACCO applies cross-modal masked reconstruction of compositional concepts with inter- and intra-modal auxiliary objectives to improve visio-linguistic compositionality in VLMs.