MMDisCo: Multi-Modal Discriminator-Guided Cooperative Diffusion for Joint Audio and Video Generation

TL;DR

This paper introduces MMDisCo, a novel method that guides pre-trained audio and video diffusion models to generate well-aligned multimodal samples efficiently, using a discriminator-based joint guidance module.

Contribution

It proposes a lightweight joint guidance approach leveraging a discriminator to improve multimodal alignment in audio-video generation using pre-trained models.

Findings

Improves multimodal alignment and fidelity with fewer parameters.

Achieves better results on benchmark datasets.

Efficiently leverages pre-trained single-modal models.

Abstract

This study aims to construct an audio-video generative model with minimal computational cost by leveraging pre-trained single-modal generative models for audio and video. To achieve this, we propose a novel method that guides single-modal models to cooperatively generate well-aligned samples across modalities. Specifically, given two pre-trained base diffusion models, we train a lightweight joint guidance module to adjust scores separately estimated by the base models to match the score of joint distribution over audio and video. We show that this guidance can be computed using the gradient of the optimal discriminator, which distinguishes real audio-video pairs from fake ones independently generated by the base models. Based on this analysis, we construct a joint guidance module by training this discriminator. Additionally, we adopt a loss function to stabilize the discriminator's gradient and make it work as a noise estimator, as in standard diffusion models. Empirical evaluations on several benchmark datasets demonstrate that our method improves both single-modal fidelity and multimodal alignment with relatively few parameters. The code is available at: https://github.com/SonyResearch/MMDisCo.