There are many variations to genetic algorithms. We can have two parent populations with two different fitness criteria (for example, lowest MSE and smoothness). We could impose restrictions on the mutation values to not be greater than 1 or less than -1. There are many different changes we could make, and these changes vary greatly, depending on the problem we are trying to optimize. For this contrived problem, the fitness was easily calculated, but for most genetic algorithms, calculating the fitness is an arduous task. For example, if we wanted to use a genetic algorithm to optimize the architecture of a convolutional neural network, we could have an individual be an array of parameters. The parameters could stand for the ...
There's more...
Get TensorFlow Machine Learning Cookbook - Second Edition now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
