“…In the field of RS, there have been various studies (Tintarev & Masthoff, 2007;Zhang & Chen, 2018) showing that equipping recommendations with personalized explanations helps to improve the system's persuasiveness and the users trust in it. In recent years, collaborative filtering methods, in particular those utilizing deep learning models (Hartford et al, 2018;Berg et al, 2017;Zheng et al, 2016), have improved recommendation performances, yet, the black box nature of deep learning models and the latent space used in collaborative filtering, make these models hard to explain.…”