Recommendation systems predict desirable recommendations for users based on their past interactions (e.g., ratings, likes, and shopping history). These systems face challenges such as data insufficiency and lack of transparency. Leveraging user reviews has been proposed as a solution to address these challenges. First of all, reviews convey user preferences so that they can be used as supplementary source of data. Secondly, reviews are a great language source to generate textual explanations, helping overcome transparency issues. Recently, due to the potential threats of AI to society, the concept of Responsible AI (RAI) has gained considerable attention. Fundamentally, RAI addresses ethical and societal concerns. Transparency is known as one of the primary principles of RAI because it helps end-users to understand the reasons behind the predictions of AI models (e.g., recommender systems). Therefore, e-commerce platforms that use recommender systems should provide users with explainable recommendations to act responsibly. Text is a prevalent style of explanation employed in a wide range of explainable recommendation systems. However, previous works often neglect the writing style of users, which can impact the effectiveness of explanations. The main objective of the paper is to develop a novel framework for generating explainable recommendations that incorporate user writing styles and preferences in the explanation generation process.
recommender system- responsible AI - multi-task learning - explanation generation - explainable recommendation - text reconstruction