Reproduced: Distribution-based Learnable Filters with Side Information for Sequential Recommendation

Sequential Recommendation aims to recommend relevant items to users based on their past interactions. In the paper "Distribution-based Learnable Filters with Side Information for Sequential Recommendation", the authors present DLFS-Rec, an innovative model for Sequential Recommendation that aims to outperform state-of-the-art models by utilizing gaussian embedding, including relevant side information as well as implementing learnable filters on the frequency domain to attenuate noise. In the following, we conduct experiments with the presented model to find out whether or not we can confirm the claimed superiority of DLFS-Rec over other state-of-the-art approaches. We find out that the papers results can be replicated as DLFS-Rec outperforms baseline models on the datasets used in the original experiments as well as on different datasets we introduce.