Taboola Algo Winter Talks: Deep Modeling for Content Recommendation // Aviv Rotman

Deep Modeling for Content Recommendation // Aviv Rotman

Taboola is powered by a recommendation engine that aims to match users with content that suits them most out of a pool of over a million possible recommendations. Deep Learning models have been gaining increasing attention in the recommendation systems community, replacing some of the traditional methods. The sparse nature of the problems in this domain and the different types of inputs offer unique challenges for feature engineering and architecture planning, in order to balance between memorization and generalization.

In this talk, we will discuss our journey to apply DNN modeling techniques for the purpose of predicting click through rate in our content recommendation system. We will deep dive into some common (and some not so common) architectures and discuss how they come into play in our problem. Specifically, we will talk about building neural networks with multiple input types (click history, text and pictures); feature engineering in the deep learning era; Tradeoffs between deep models, shallow models and the combination of the two; and other tips regarding network architectures.