Experimental Design - Graph Trend Filtering Networks for Recommendation

We choose the research paper Graph Trend Filtering Networks for Recommendation because
we found this topic interesting, the paper was understandable for us and also there was a GitHub
repository provided with the source, leading us to believe that reproducibility would be easy.
This paper expanded on recommender systems and offered a novel solution for this field. The key role
of recommender systems is to predict whether users are likely to interact with items (e.g. products,
songs, movies, etc.) based on their previous interactions, including clicks, add-to-cart, purchases, and
so on. The research of the author also briefly went through what collaborative filtering (CF) techniques
are. These techniques are developed to model user-item interactions on historical view, assuming that
users who behave similarly are likely to have similar preferences towards items. The paper utilized this
with neural networks to create neural collaborative filtering specifically their version of graph neural
networks (GNNs), which is graph trend filtering networks for recommendations (GTN)