Friday, 17th of January11h (room R2014, 660 building) (see location)
A journey to causal advertisingIn a culture where claims are backed with data, digital advertising shall demonstrate and optimize its causal effect. Measuring the causal effect of ads is essential but difficult. In addition, measuring it is not enough we would like to optimize it or at least to increase it. We will present what is a causal advertising and how it differs from regular advertising. Then will present two approaches, one based on counterfactual estimation and one based on uplift prediction. Our studies are based on interventional data, some of those data are available as open Criteo datasets.
All TAU seminars: here