Vision-Language Models (VLMs) are typically trained on a diverse set of multi-modal domains, yet current practices rely on costly manual tuning. We propose MaD-Mix, a principled and computationally efficient framework that derives multi-modal data mixtures for VLM training. MaD-Mix formulates data mixing as modality-aware domain alignment maximization and obtains closed-form multi-modal alignment scores from the Fenchel dual through inter-modal coupling variables. MaD-Mix systematically handles domains with missing modalities, allowing for the integration of language-only domains. Empirical evaluations across 0.5B and 7B models demonstrate that MaD-Mix accelerates VLM training across diverse benchmarks. MaD-Mix matches human-tuned data mixtures using 22% fewer training steps in image-text instruction tuning. In complex tri-modal video-image-text scenarios, where manual tuning becomes impractical, MaD-Mix boosts average accuracy over uniform weights, with negligible mixture computation overhead (< 1 GPU-hour), enabling scalable mixture design for modern VLM pipelines.
@article{arxiv.2602.07790,
title = {MaD-Mix: Multi-Modal Data Mixtures via Latent Space Coupling for Vision-Language Model Training},
author = {Wanyun Xie and Francesco Tonin and Volkan Cevher},
journal= {arXiv preprint arXiv:2602.07790},
year = {2026}
}