Budget Allocation as a Multi-Agent System of Contextual & Continuous Bandits
Benjamin Han,Carl Arndt
Budget allocation for online advertising suffers from multiple complications, including significant delay between the initial ad impression to the call to action as well as cold-start prediction problems for ad campaigns with limited or no historical performance data. To address these issues, we introduce the Contextual Budgeting System (CBS ), a budget allocation framework using a multi-agent system of contextual & continuous Multi-Armed Bandits. Our proposed solution decomposes the problem into a convex optimization problem whose objective is drawn using Thompson Sampling. In order to efficiently deal with context and cold-start, we propose a transfer learning mechanism using supervised learning methods that augment simple parametric models.We apply an implementation of this algorithm to all spending for new driver acquisition at Lyft and measure a (22 u00b1 10)% improvement in the mean Cost Per user Acquisition (CPA) over a previous non-contextual model, based on Markov Chain Monte-Carlo, generating tens of millions of dollars annually in efficiency improvements while also increasing total user acquisition.
Discussion
-
https://evolution.org.ua/Please let me know if you're looking for a article writer for your site. You have some really great articles and I feel I would be a good asset. If you ever want to take skme oof the load off, I'd love to write some content for your blog inn exchange for a link bac to mine. Please blast me an email if interested. Many thanks!
Reply