Machine learning for targeted display advertising: transfer learning in action

Abstract

This paper presents a detailed discussion of problem formulation and data representation issues in the design, deployment, and operation of a massive-scale machine learning system for targeted display advertising. Notably, the machine learning system itself is deployed and has been in continual use for years, for thousands of advertising campaigns (in contrast to simply having the models from the system be deployed). In this application, acquiring sufficient data for training from the ideal sampling distribution is prohibitively expensive. Instead, data are drawn from surrogate domains and learning tasks, and then transferred to the target task. We present the design of this multistage transfer learning system, highlighting the problem formulation aspects. We then present a detailed experimental evaluation, showing that the different transfer stages indeed each add value. We next present production results across a variety of advertising clients from a variety of industries, illustrating the performance of the system in use. We close the paper with a collection of lessons learned from the work over half a decade on this complex, deployed, and broadly used machine learning system.

A survey of transfer learning

“…Online display web advertising is a growing industry where transfer learning is used to optimally predict targeted ads. In the paper by Perlich [90], a transfer learning approach is employed that uses the weighted outputs of multiple source classifiers to enhance a target classifier trained to predict targeted online display advertising results. The paper by Kan [56] addresses the field of facial recognition and is able to use face image information from one ethnic group to improve the learning of a classifier for a different ethnic group.…”

Bigger is Better, but at What Cost?Estimating the Economic Value of Incremental Data Assets

“…19 The company uses both first-party and thirdparty user behavior data to match the right ads to the right users. As is common in the ad tech industry, Dstillery has its own native data but also has access to data segments sold by third parties.…”

Machine learning for science and society

“…We accepted a set of papers that addresses several critical aspects of society: these papers may influence or are influencing the way we perform repairs on fundamental infrastructure (Li et al 2013), the way we predict severe weather and aviation turbulence (McGovern et al 2013;Williams 2013), how we conduct tax audits (Kong and Saar-Tsechansky 2013), whether we can detect privacy breaches in access to healthcare data (Menon et al 2013), how we target advertisements (Perlich et al 2013), and how we link census datasets to track an individual's career trajectory (Antonie et al 2013).…”

“…In particular, they needed to explicitly not use certain information, which could inadvertently lead to a bias in subsequent studies involving their linkage system. Another perspective comes from Perlich et al (2013), who discuss that data from outside the exact target domain can be useful. Specifically, they state that "acquiring a large amount of data that is not from the optimal data generating distribution can be better than acquiring only a small amount of data from the optimal data generating distribution.…”