Can I read Visualising Basins of Attraction For The Cross-Entropy and The Squared Error Neural Network Loss Functions - 1-S2.0-S0925231220303593-Main on EtoBox?

Visualising Basins of Attraction For The Cross-Entropy and The Squared Error Neural Network Loss Functions - 1-S2.0-S0925231220303593-Main by Kauan is a document available to read on EtoBox.

What is Visualising Basins of Attraction For The Cross-Entropy and The Squared Error Neural Network Loss Functions - 1-S2.0-S0925231220303593-Main about?

This paper presents a novel method for visualizing the basins of attraction and stationary points of neural network loss functions, specifically comparing quadratic loss and entropic loss. The empirical study confirms that entropic loss has stronger gradients and fewer stationary points, making it more favorable for gradient-based training, while quadratic loss is more resilient to overfitting. The findings contribute to a better understanding of the error landscapes in neural networks, which is crucial for

Author

Kauan

Language

EN