AUC Optimisation and Collaborative Filtering
This work addresses the need for efficient and theoretically grounded ranking optimization in collaborative filtering for recommendation systems, representing an incremental improvement over existing methods.
The paper tackles the problem of optimizing ranking performance in recommendation systems by proposing a class of differentiable objective functions that serve as smooth surrogates for the Area Under the ROC Curve (AUC), with a special case focusing on top-of-list prioritization, and demonstrates its efficacy through computational results on synthetic and real data, showing scalability with linear time complexity.
In recommendation systems, one is interested in the ranking of the predicted items as opposed to other losses such as the mean squared error. Although a variety of ways to evaluate rankings exist in the literature, here we focus on the Area Under the ROC Curve (AUC) as it widely used and has a strong theoretical underpinning. In practical recommendation, only items at the top of the ranked list are presented to the users. With this in mind, we propose a class of objective functions over matrix factorisations which primarily represent a smooth surrogate for the real AUC, and in a special case we show how to prioritise the top of the list. The objectives are differentiable and optimised through a carefully designed stochastic gradient-descent-based algorithm which scales linearly with the size of the data. In the special case of square loss we show how to improve computational complexity by leveraging previously computed measures. To understand theoretically the underlying matrix factorisation approaches we study both the consistency of the loss functions with respect to AUC, and generalisation using Rademacher theory. The resulting generalisation analysis gives strong motivation for the optimisation under study. Finally, we provide computation results as to the efficacy of the proposed method using synthetic and real data.