HIGHLIGHTS
- who: Input2 from the (UNIVERSITY) have published the research: Loss Function Dynamics and Landscape for Deep Neural Networks Trained with Quadratic Loss, in the Journal: (JOURNAL)
- what: In this paper the dynamics of learning and the surface of the standard cross-entropy loss function and the currently popular mean squared error (MSE) loss function for scale-invariant networks with normalization are studied. The authors propose normalizing the weights at the current step. To analyze the first phase, the authors investigate the linear connectivity of models (mode connectivity). Now, the authors examine the behavior of networks . . .
If you want to have access to all the content you need to log in!
Thanks :)
If you don't have an account, you can create one here.