Why Relational Learning Matters - Automated Feature Engineering on Relational Data and Time Series

The deep learning revolution has allowed us to automate feature engineering on unstructured data such as images, text and speech, replacing 30 years of research on manual feature engineering in these domains. However, when  data scientists work with structured data, such as relational data sets and time series, most features are still engineered manually. 

This talk will introduce algorithms and techniques that can close that gap and demonstrate how relational learning can be used to automate feature engineering and significantly reduce the time and costs of data science projects on relational data and time series.