Automatic ad format selection via contextual bandits
Authors:
Open Access
- 27 October 2013
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 1587-1594
Abstract
Visual design plays an important role in online display advertising: changing the layout of an online ad can increase or decrease its effectiveness, measured in terms of click-through rate (CTR) or total revenue. The decision of which lay- out to use for an ad involves a trade-off: using a layout provides feedback about its effectiveness (exploration), but collecting that feedback requires sacrificing the immediate reward of using a layout we already know is effective (exploitation). To balance exploration with exploitation, we pose automatic layout selection as a contextual bandit problem. There are many bandit algorithms, each generating a policy which must be evaluated. It is impractical to test each policy on live traffic. However, we have found that offline replay (a.k.a. exploration scavenging) can be adapted to provide an accurate estimator for the performance of ad layout policies at Linkedin, using only historical data about the effectiveness of layouts. We describe the development of our offline replayer, and benchmark a number of common bandit algorithms.Keywords
This publication has 13 references indexed in Scilit:
- Ad click prediction2013
- Dynamic ad layout revenue optimization for display advertising2012
- Unbiased offline evaluation of contextual-bandit-based news article recommendation algorithms2011
- Spatio-temporal models for estimating click-through rate2009
- Exploration–exploitation tradeoff using variance estimates in multi-armed banditsTheoretical Computer Science, 2009
- Exploration scavenging2008
- Predicting clicks2007
- Multi-armed Bandit Algorithms and Empirical Evaluation2005
- Finite-time Analysis of the Multiarmed Bandit ProblemMachine Learning, 2002
- Reinforcement Learning: An IntroductionIEEE Transactions on Neural Networks, 1998