TY  - RPRT
TI  - VCMM: Variance-Calibrated Momentum for Multimodal Learning
AU  - Zhongjing Gu
AU  - Chenyang Huang
AU  - Yufa Feng
AU  - Chong He
AU  - Qinxu Ding
AU  - Yiming Cui
PY  - 2026
UR  - https://arxiv.org/abs/2609.27577
ID  - 2609.27577
ER  -