Deep Learning Techniques for Large Scale Multimodal Click-Through Rate (CTR) Prediction and Search Ranking in e-commerce
 

Click-Through Rate (CTR) prediction is an integral part of online search advertising systems where it is utilized as an input to auctions which determine the final ranking of products for a particular user and for each query. In this talk, we introduce an ensemble learning approach for CTR prediction which is based on historical or behavioral signals for older products, as well as content-based features for new products. We obtain representations from texts and images by utilizing state-of-the-art deep learning techniques and employ multimodal learning to combine these different signals. Moreover, we introduce a multimodal learning to rank model that combines traditional text features with visual semantic features transferred from a deep convolutional neural network.


 

Kamelia Aryafar, Ph.D. is the Director of Machine Learning (ML) at Overstock.com. She leads a team of data scientists and machine learning engineers focused on building scalable ML/AI and computer vision tools to curate a personalized experience for Overstock users. Prior to Overstock she has been a Senior Data Scientist at Etsy for more than four years. Before Etsy, she was doing a Ph.D. in computer science and machine learning in Drexel University, building large-scale music classification models.