AutoML Toolkit Deep Dive - Automating Feature Engineering and Model Optimization

Dive deep into the AutoML Toolkit in this hour-long conference talk from Databricks. Explore how Databricks Labs is automating and accelerating feature engineering, feature importance selection, model selection and tuning, model serving/deployment, model documentation with MLflow, and inference and scoring. Learn about the toolkit's history, features, models, overrides, Python workflows, and practical applications using a wine dataset. Discover how to set up the toolkit, perform feature engineering and optimization, analyze results, and execute manual runs. Gain insights into classification, case sampling, feature interactions, and upcoming features. Understand pipeline model instrumentation, prediction, and the feature engineering pipeline to streamline your machine learning workflows and boost productivity.

Intro
History
Features
Models
Overrides
Python
Workflows
Wine Dataset
Setup
Feature Engineering
Feature Optimization
Results
Manual Run
Recap
Introduction
Classification
Case Sampling
Feature Interactions
Feature Modes
Upcoming Features
Getting Started
Pipeline Model Instrumentation
Pipeline Model Prediction
Feature Engineering Pipeline