Auto ML Is More Than Hyperparameter Optimization or Neural Architecture Search

The field of auto ML is much, much broader than hyperparameter optimization or neural architecture search. Everything from collecting my raw data to deploying a model and monitoring it should fall under the umbrella of his broader auto ML problem. And so in some sense you can certainly use the product without doing hyperparameter Optimizer or Neural Architecture Search but we kind of think that everything from beginning to end should be more automated. Does that make sense?