Zero-Shot Recommender Systems

HAO DING, Yifei Ma, Anoop Deoras, Bernie Wang, Hao Wang

Keywords: Zero Shot Learning, Recommender Systems, Neural Networks, Bayesian

Abstract: Performance of recommender systems (RecSys) relies heavily on the amount of training data available. This poses a chicken-and-egg problem for early-stage products, whose amount of data, in turn, relies on the performance of their RecSys. In this paper, we explore the possibility of zero-shot learning in RecSys, to enable generalization from an old dataset to an entirely new dataset. We develop an algorithm, dubbed ZEro-Shot Recommenders (ZESRec), that is trained on an old dataset and generalize to a new one where there are neither overlapping users nor overlapping items, a setting that contrasts typical cross-domain RecSys that has either overlapping users or items. Different from previous methods that use categorical item indices (i.e., item ID), ZESRec uses items' generic features, such as natural-language descriptions, product images, and videos, as their continuous indices, and therefore naturally generalizes to any unseen items. In terms of users, ZESRec builds upon recent advances on sequential RecSys to represent users using their interactions with items, thereby generalizing to unseen users as well. We study three pairs of real-world RecSys datasets and demonstrate that ZESRec can successfully enable recommendations in such a zero-shot setting, opening up new opportunities for resolving the chicken-and-egg problem for data-scarce startups or early-stage products.

One-sentence Summary: A novel hierarchical Bayesian model that performs zero-shot recommendation in a target domain where there are neither overlapping users nor overlapping items with the source domain.