Scalable AutoML for Time Series Forecasting Using Ray

Explore scalable AutoML techniques for time series forecasting using Ray in this 22-minute conference talk from OpML '20. Dive into the development of an easy-to-use toolkit that leverages machine learning and deep learning methods to outperform traditional forecasting approaches. Learn how the speakers built an AutoML toolkit on top of Ray, automating feature generation, selection, model selection, and hyper-parameter tuning in a distributed manner. Gain insights into real-world applications, including network quality analysis, log analysis for data center operations, and predictive maintenance. Discover the toolkit's architecture, core components, ML framework, and software stack. Follow the training workflow and examine a reference use case with project background and examples. Conclude with a summary of key takeaways and future work in the field of AutoML for time series forecasting.

Introduction
Background
Time Series
Ray
Core Parts
ML Framework
Software Stack
Training Workflow
Recipe
New Project
Reference Use Case
Project Background
Project Example
Real Case
Summary
Future work