LayerLock: Non-collapsing Representation Learning with Progressive FreezingGoker Erdogan, Nikhil Parthasarathy, Catalin Ionescu, Drew Hudson, Alexander Lerchner, Andrew Zisserman, Mehdi Sajjadi, Joao Carreirahttps://arxiv.org/abs/2509.10156
LayerLock: Non-collapsing Representation Learning with Progressive FreezingWe introduce LayerLock, a simple yet effective approach for self-supervised visual representation learning, that gradually transitions from pixel to latent prediction through progressive layer freezing. First, we make the observation that during training of video masked-autoencoding (MAE) models, ViT layers converge in the order of their depth: shallower layers converge early, deeper layers converge late. We then show that this observation can be exploited to accelerate standard MAE by progress…