Causal Inference Methods and Applications in Tech
Friday, Jan. 6, 2023 8:00 AM - 10:00 AM (CST)
- Chair: Susan Athey, Stanford University
Double Robust Causal Effect Extrapolation w/ Applications at Netflix
AbstractBeing able to extrapolate the causal effect beyond the study population to a given new population allows us to increase the speed and utility of our learnings. Building upon existing literature, we developed a framework and software tool for causal effect generalization and transportation. Generalizability addresses the problem of extrapolating the causal effect from the study population to a larger target population that contains it. Transportability addresses the problem of extrapolating the causal effect from a study population to a different target population. We describe some Netflix applications of this framework.
The Science of Pricing Experimentation at Amazon
AbstractIn order to improve prices at Amazon, we created Pricing Labs, a price experimentation platform. Since we do not price discriminate, we must run product-randomized experiments. In this presentation, I will discuss how we randomize to prevent spillovers, run different experimental designs (i.e. crossovers) to improve precision, and control for demand trends and differences in treatment groups to get more precise treatment effect estimates.
Estimating the Long-Term Effects of Novel Treatments
AbstractPolicy makers typically face the problem of wanting to estimate the long-term effects of novel treatments, while only having historical data of older treatment options. We assume access to a long-term dataset where only past treatments were administered and a short-term dataset where novel treatments have been administered. We propose a surrogate based approach where we assume that the long-term effect is channeled through a multitude of available short-term proxies. Our work combines three major recent techniques in the causal machine learning literature: surrogate indices, dynamic treatment effect estimation and double machine learning, in a unified pipeline. We show that our method is consistent and provides root-n asymptotically normal estimates under a Markovian assumption on the data and the observational policy. We use a data-set from a major corporation that includes customer investments over a three year period to create a semi-synthetic data distribution where the major qualitative properties of the real dataset are preserved. We evaluate the performance of our method and discuss practical challenges of deploying our formal methodology and how to address them.
- C1 - Econometric and Statistical Methods and Methodology: General
- M2 - Business Economics