Background:
Open Benchmarking for Click-Through Rate Prediction https://arxiv.org/pdf/2009.05794.pdf
In particular, we will explore how DSPs (demand side platforms) optimize ad campaigns for advertisers.
The focus will be on how CTR (click-through-rates) are optimized via real-time-bidding (RTB) models. We will discuss modeling challenges such as how best to handle categorical or imbalanced training data. You will come away with a deeper understanding of the system design trade-offs and be able to deploy models at scale in your own work.