Automatic Debiased Machine Learning via Neural Nets
Debiased machine learning based on regression requires estimation of an unknown debiasing function. We give an automatic method of estimating the debiasing function based on any machine learner. This method only depends on the parameter of interest and the machine learner. We give primitive regularity conditions for neural net estimation of the debiasing function and show its superior performance relative to inverse probability weighting in estimating the average treatment effect in a state of the art example.
Date: 23 May 2023, 13:00 (Tuesday, 5th week, Trinity 2023)
Venue: Manor Road Building, Manor Road OX1 3UQ
Venue Details: Skills Lab or https://zoom.us/j/93054414699?pwd=NEFiL2ZNc0t5N3ZIUTE2VEh5OXhZUT09
Speaker: Whitney Newey (MIT)
Organising department: Department of Economics
Part of: Nuffield Econometrics Seminar
Booking required?: Not required
Audience: Members of the University only
Editors: Emma Heritage, Daria Ihnatenko