VAE-IPS: A Deep Generative Recommendation Method for Unbiased Learning from Implicit Feedback
Shashank Gupta, Harrie Oosterhuis, and Maarten de Rijke Published in CONSEQUENCES+REVEAL Workshop at RecSys ’22, 2022. [pdf]
Variational autoencoders (VAEs) are the state-of-the-art model for recommendation with implicit feedback signals. Unfortunately, implicit feedback suffers from selection bias, e.g., popularity bias, position bias, etc., and as a result, training from such signals produces biased recommendation models. Existing methods for debiasing the learning process have not been applied in a generative setting. In this work, we address this gap by introducing an inverse propensity scoring (IPS) based unbiased training method for VAEs from implicit feedback data, VAE-IPS, which is provably unbiased w.r.t. selection bias. Our experimental results show that the proposed VAE-IPS model reaches significantly higher performance than existing baselines.
Recommended citation:
S. Gupta, H. Oosterhuis, and M. de Rijke. "VAE-IPS: A Deep Generative Recommendation Method for Unbiased Learning from Implicit Feedback." CONSEQUENCES+REVEAL Workshop at RecSys ’22, September 23, 2022, Seattle, USA.