A Type System for Feature Engineering in Machine Learning

A Type System for Feature Engineering in Machine Learning

Explore a 22-minute conference talk from ACM SIGPLAN's TyDe'23 that delves into the application of Martin-Löf's type theory in feature engineering. Learn how Weixi Ma, Serena Chan, and Fei Yu demonstrate the dependent type system of F3 (Facebook Feature Framework) and its practical applications in daily feature engineering. Gain insights into how a compiler for feature engineering can benefit from type theory implementation techniques, and understand the process of formalizing domain knowledge into data operations for converting raw data into proper machine learning model inputs.