Elliptic Loss Regularization

Ali Hasan, Haoming Yang, Yuting Ng +1

Regularizing neural networks is important for anticipating model behavior in regions of the data space that are not well represented. In this work, we propose a regularization tech…