Diffusion Models as Masked Audio-Video Learners
This work addresses efficiency improvements for researchers and practitioners using audio-visual learning models, but it is incremental as it builds directly on the existing MAViL framework.
The paper tackled the problem of improving the efficiency of the MAViL audio-video pre-training framework by integrating diffusion models, resulting in a 32% reduction in pre-training FLOPS and an 18% decrease in wall clock time without compromising downstream audio-classification performance.
Over the past several years, the synchronization between audio and visual signals has been leveraged to learn richer audio-visual representations. Aided by the large availability of unlabeled videos, many unsupervised training frameworks have demonstrated impressive results in various downstream audio and video tasks. Recently, Masked Audio-Video Learners (MAViL) has emerged as a state-of-the-art audio-video pre-training framework. MAViL couples contrastive learning with masked autoencoding to jointly reconstruct audio spectrograms and video frames by fusing information from both modalities. In this paper, we study the potential synergy between diffusion models and MAViL, seeking to derive mutual benefits from these two frameworks. The incorporation of diffusion into MAViL, combined with various training efficiency methodologies that include the utilization of a masking ratio curriculum and adaptive batch sizing, results in a notable 32% reduction in pre-training Floating-Point Operations (FLOPS) and an 18% decrease in pre-training wall clock time. Crucially, this enhanced efficiency does not compromise the model's performance in downstream audio-classification tasks when compared to MAViL's performance.