Today at AWS re:Invent in Las Vegas, the company announced AutoPilot, a tool that gives you greater visibility in automated machine learning model creation. This is available as part of the new SageMaker Studio also announced today.
As AWS CEO Andy Jassy pointed out on stage today, one of the problems with AutoML is that it’s basically a black box. “First they build this OK, simple model initially, but that is a total black box. If you want to improve a mediocre model or just evolve it for your business, you you have no idea how it was built.” he explained.
The idea behind AutoPilot is to give you the ease of model creation you get from and AutoML-generated model, but also giving you much deeper insight into how the system built the model. “AutoMl in a way to create a model automatically but give full visibility and control,” Jassy said.
You can look at the model’s parameters, and see 50 automated model, then the it provides you with a leader board of what models were the best. You can look at the notebook, and also see what trade-offs were made to generate that best model. It may be the most accurate, but sacrifices speed to get that.
Your company may have its own set of unique requirements and you can choose the best model based on whatever parameters you consider to be most important for your company, even though this was generated in an automated fashion.
With this insight and visibility into your automatically generated model, you can pick the best one based on you requirements. Once you have the model, you like best, you can go into SageMaker Studio, select it and launch it with a single click.