Mastering Deep Learning: Designing Loss Functions

In past articles, we have seen how neural networks approximate non-linearities and the reason for their extraordinary efficiency. In this article, let us expand our understanding by exploring loss functions. Firstly, what exactly is a loss function, and what purpose does it serve? To answer our question, let us recall how we represent neural networks. A simple neural network that maps a set of inputs X to a set of outputs Y is represented as follows: <a href="https://pub.towardsai.net/mastering-deep-learning-designing-loss-functions-e12a68c5ccf9">Website</a>