The Present and Future of Machine Learning for Ad Creatives

Abstract

In display advertisements, ad creatives, such as images and texts, play an essential role in delivering product information to customers efficiently. Such ad creatives are mainly created and operated in large quantities manually. However, creating effective ad creatives is generally difficult and extremely expensive. Therefore, to create and operate ad creatives with high serving effects, research, and development based on the machine learning techniques are actively performed worldwide. In particular, the following tasks are recently gathering attention: predicting CTR/CVR for ad creative before serving to consumers, analyzing the elements that make up high-performing ad creatives, and generating effective ad creatives. In this talk, I will introduce my research topic: the framework of ad creative evaluation based on CVR prediction to support the creation of effective ad creative. This research is presented at KDD2019, one of the most famous conferences in data mining. Also, I will introduce several research topics related to the latest ad technology and discuss the current status and prospects achieved by the research on ad creative and machine learning techniques.