CaReFlow: Cyclic Adaptive Rectified Flow for Multimodal Fusion

TL;DR

This paper introduces CaReFlow, a novel multimodal fusion method that uses cyclic adaptive rectified flow to better align modality distributions, reducing the modality gap and improving performance in affective computing tasks.

Contribution

It extends rectified flow for one-to-many distribution mapping and proposes adaptive relaxed alignment to enhance multimodal feature alignment.

Findings

Achieves competitive results on multimodal affective computing tasks.

Effectively reduces modality gap as shown by visualizations.

Enables robust distribution transformation with limited paired data.

Abstract

Modality gap significantly restricts the effectiveness of multimodal fusion. Previous methods often use techniques such as diffusion models and adversarial learning to reduce the modality gap, but they typically focus on one-to-one alignment without exposing the data points of the source modality to the global distribution information of the target modality. To this end, leveraging the characteristic of rectified flow that can map one distribution to another via a straight trajectory, we extend rectified flow for modality distribution mapping. Specifically, we leverage the `one-to-many mapping' strategy in rectified flow that allows each data point of the source modality to observe the overall target distribution. This also alleviates the issue of insufficient paired data within each sample, enabling a more robust distribution transformation. Moreover, to achieve more accurate distribution mapping and address the ambiguous flow directions in one-to-many mapping, we design `adaptive relaxed alignment', enforcing stricter alignment for modality pairs belonging to the same sample, while applying relaxed mapping for pairs not belonging to the same sample or category. Additionally, to prevent information loss during distribution mapping, we introduce `cyclic rectified flow' to ensure the transferred features can be translated back to the original features, allowing multimodal representations to learn sufficient modality-specific information. After distribution alignment, our approach achieves very competitive results on multiple tasks of multimodal affective computing even with a simple fusion method, and visualizations verify that it can effectively reduce the modality gap.