The following slides are an overview of AutoML. This is an updated version of the slides presented at SuperComputing18. Additionally, this session covers an introduction to H2O for model selection and Comet.ml for hyperparameter optimization.
Introduction to AutoML
H2O
H2O is a tool that allows you to perform Automated Machine Learning. A Jupyter notebook with an introduction to H2O can be found in the GitHub repository. The binder path to the repository is located here
Comet.ml
Comet.ml is a workflow management tool for Machine Learning. However, it also has the ability to automate hyperparameter optimization particularly for Neural Networks. I cover that here.
Automated provisioning of VMs with Ansible
For this session, I will be using Genesis Cloud for demonstrating AutoML with H2O. An ansible playbook for setting up H2O in a conda environment is provided in this GitHub repository. This sets up a JupyterLab for the API usage as well as giving access to the H2O web interface.