TensorFlow Machine Learning Projects

Bayes' rule in neural networks

Traditionally, neural networks produce a point estimate by optimizing weights and biases to minimize a loss function, such as the mean squared error in regression problems. As mentioned earlier, this is similar to finding parameters using the Maximum likelihood estimation criteria:

Typically, we obtain the best parameters through backpropagation in neural networks. To avoid overfitting, we introduce a regularizer of  norm over weights. If you are not aware of regularization, please refer to the following Andrew ...