Large Scale Co-Regularized Ranking
Evgeni Tsivtsivadze, Katja Hofmann, Tom Heskes, J. Fürnkranz, E. Hüllermeyer · UvA-DARE (University of Amsterdam) · 2012
Abstract. As unlabeled data is usually easy to collect, semi-supervised learning algorithms that can be trained on large amounts of unlabeled and labeled data are becoming increas-ingly popular for ranking and preference learning problems [6, 23, 8, 21]. However, the computational complexity of the vast majority of these (pairwise) ranking and preference learn-ing methods is super-linear, as optimizing an objective func-tion over all possible pairs of data points is computationally expensive. This paper builds upon [16] and proposes a novel large scale co-regularized algorithm that can take unlabeled data into account. This algorithm is suitable for learning to rank when large amounts of labeled and unlabeled data are available for training. Most importantly, the complexity of our algorithm does not depend on the size of the dataset. We evaluate the proposed algorithm using several publicly available datasets from the information retrieval (IR) domain, and show that it improves performance over supervised methods. Finally, we discuss possible implications of our algorithm for learn-ing with implicit feedback in an online setting. 1 Introduction and