Bibliography (3):

1. MAE: Masked Autoencoders Are Scalable Vision Learners

2. Attention Is All You Need

3. https://github.com/facebookresearch/AudioMAE