Due to the advance of big data and increasing availability of smartphones equipped with various sensors and networking capability, it provides new opportunities for innovating mobile application advertising services. Unlike Web, where cookies for identifying users store in web browsers, there is a challenge for mobile apps to track users so acquiring sufficient data for training from ideal sampling distribution is computationally and economically expensive. Thus, a massive-scale intelligent system for targeted mobile app display advertising is developed to capture data for the learning task and transfer extracted knowledge back to the target task. The proposed system is evaluated by real world data and deployed to an advertising agency, illustrating the practicability and applicability.
Proceeding of Conference on Technologies and Applications of Artificial Intelligence (TAAI 2013)=第十八屆人工智慧與應用學術研討會論文集, pp. 353-354