Data Science

AI Improves the Frequency and Quality of Mobile App Notifications

Researchers from Leopard Mobile, Taiwan’s first mobile internet company, recently developed a deep learning recommendation system that can improve smartphone notifications and pop-up advertisements. One of the big challenges developers and advertisers face is knowing when to deploy push notifications. If you send too many, people may delete your app, send too few,  people might forget your app even exists.
“Using push notification service properly can increase the using rate of Apps and further increase the number of advertisement display,” the researchers explained in their paper. “It is important to forecast users’ preference and frequency on push notifications precisely,” they explained.
Using NVIDIA GeForce GTX 1080 GPUs and the cuDNN-accelerated TensorFlow deep learning framework, the researchers trained their deep neural network on browsing history, shopping history, and financial details, provided to them by Leopard Mobile Inc. and Cheetah Mobile Inc, of over 623 million users.
Once trained, the team used NVIDIA GPUs in the cloud to execute the model, which resulted in more personalized notifications and advertisements.

The researcher’s architecture and flowchart.

“The results showed that our system effectively decreased the number of popups and increased the click-through rate and 7-day retention rate,” the researchers wrote. “We expect to provide more convenient user scenarios to end users or enterprises.”
The team is looking to improve their deep learning model. By doing so, they hope to decrease complicated tasks and to train a high-performance advertisement recommendation system.
Read More >

Discuss (0)