Self-Supervised Learning for Covariance Estimation

arXiv:2403.08662v1 Announce Type: cross
Abstract: We consider the use of deep learning for covariance estimation. We propose to globally learn a neural network that will then be applied locally at inference time. Leveraging recent advancements in self-supervised foundational models, we train the network without any labeling by simply masking different samples and learning to predict their covariance given their surrounding neighbors. The architecture is based on the popular attention mechanism. Its main advantage over classical methods is the automatic exploitation …

abstract arxiv covariance cs.lg deep learning eess.sp foundational models inference labeling learn masking network neural network samples self-supervised learning supervised learning train type will