InfMasking: Unleashing Synergistic Information by Contrastive Multimodal Interactions

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wen, Liangjian, Dai, Qun, Liu, Jianzhuang, Zheng, Jiangtao, Dai, Yong, Wang, Dongkai, Kang, Zhao, Wang, Jun, Xu, Zenglin, Duan, Jiang
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917182087102464
author Wen, Liangjian
Dai, Qun
Liu, Jianzhuang
Zheng, Jiangtao
Dai, Yong
Wang, Dongkai
Kang, Zhao
Wang, Jun
Xu, Zenglin
Duan, Jiang
author_facet Wen, Liangjian
Dai, Qun
Liu, Jianzhuang
Zheng, Jiangtao
Dai, Yong
Wang, Dongkai
Kang, Zhao
Wang, Jun
Xu, Zenglin
Duan, Jiang
contents In multimodal representation learning, synergistic interactions between modalities not only provide complementary information but also create unique outcomes through specific interaction patterns that no single modality could achieve alone. Existing methods may struggle to effectively capture the full spectrum of synergistic information, leading to suboptimal performance in tasks where such interactions are critical. This is particularly problematic because synergistic information constitutes the fundamental value proposition of multimodal representation. To address this challenge, we introduce InfMasking, a contrastive synergistic information extraction method designed to enhance synergistic information through an Infinite Masking strategy. InfMasking stochastically occludes most features from each modality during fusion, preserving only partial information to create representations with varied synergistic patterns. Unmasked fused representations are then aligned with masked ones through mutual information maximization to encode comprehensive synergistic information. This infinite masking strategy enables capturing richer interactions by exposing the model to diverse partial modality combinations during training. As computing mutual information estimates with infinite masking is computationally prohibitive, we derive an InfMasking loss to approximate this calculation. Through controlled experiments, we demonstrate that InfMasking effectively enhances synergistic information between modalities. In evaluations on large-scale real-world datasets, InfMasking achieves state-of-the-art performance across seven benchmarks. Code is released at https://github.com/brightest66/InfMasking.
format Preprint
id arxiv_https___arxiv_org_abs_2509_25270
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle InfMasking: Unleashing Synergistic Information by Contrastive Multimodal Interactions
Wen, Liangjian
Dai, Qun
Liu, Jianzhuang
Zheng, Jiangtao
Dai, Yong
Wang, Dongkai
Kang, Zhao
Wang, Jun
Xu, Zenglin
Duan, Jiang
Machine Learning
Artificial Intelligence
Computer Vision and Pattern Recognition
In multimodal representation learning, synergistic interactions between modalities not only provide complementary information but also create unique outcomes through specific interaction patterns that no single modality could achieve alone. Existing methods may struggle to effectively capture the full spectrum of synergistic information, leading to suboptimal performance in tasks where such interactions are critical. This is particularly problematic because synergistic information constitutes the fundamental value proposition of multimodal representation. To address this challenge, we introduce InfMasking, a contrastive synergistic information extraction method designed to enhance synergistic information through an Infinite Masking strategy. InfMasking stochastically occludes most features from each modality during fusion, preserving only partial information to create representations with varied synergistic patterns. Unmasked fused representations are then aligned with masked ones through mutual information maximization to encode comprehensive synergistic information. This infinite masking strategy enables capturing richer interactions by exposing the model to diverse partial modality combinations during training. As computing mutual information estimates with infinite masking is computationally prohibitive, we derive an InfMasking loss to approximate this calculation. Through controlled experiments, we demonstrate that InfMasking effectively enhances synergistic information between modalities. In evaluations on large-scale real-world datasets, InfMasking achieves state-of-the-art performance across seven benchmarks. Code is released at https://github.com/brightest66/InfMasking.
title InfMasking: Unleashing Synergistic Information by Contrastive Multimodal Interactions
topic Machine Learning
Artificial Intelligence
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2509.25270