A Correlation Maximization Approach for Cross Domain Co-Embeddings

TL;DR

This paper introduces ImplicitCE, a novel algorithm that creates cross-domain user and item embeddings to improve recommendations for new users by maximizing correlation between predicted and actual user affinities.

Contribution

ImplicitCE is the first end-to-end trainable model that learns cross-domain embeddings and a transformation function using correlation maximization, with a new Sample Correlation Update training method.

Findings

ImplicitCE outperforms state-of-the-art algorithms on large datasets.

Sample Correlation Update effectively trains the model with high correlation.

The approach is applicable to large-scale recommendation scenarios.

Abstract

Although modern recommendation systems can exploit the structure in users' item feedback, most are powerless in the face of new users who provide no structure for them to exploit. In this paper we introduce ImplicitCE, an algorithm for recommending items to new users during their sign-up flow. ImplicitCE works by transforming users' implicit feedback towards auxiliary domain items into an embedding in the target domain item embedding space. ImplicitCE learns these embedding spaces and transformation function in an end-to-end fashion and can co-embed users and items with any differentiable similarity function. To train ImplicitCE we explore methods for maximizing the correlations between model predictions and users' affinities and introduce Sample Correlation Update, a novel and extremely simple training strategy. Finally, we show that ImplicitCE trained with Sample Correlation Update outperforms a variety of state of the art algorithms and loss functions on both a large scale Twitter dataset and the DBLP dataset.