Analysis of LightGCN in Graph Recommendation Algorithms
In this article, we delve into the core themes of LightGCN in the context of graph recommendation algorithms. While originally designed for graph classification, LightGCN introduces a linear propagation approach for embeddings that proves to enhance performance. We replicate the original findings, investigate the robustness of LightGCN across various datasets and metrics, and also explore the potential of using Graph Diffusion as a means of augmenting signal propagation in LightGCN.
Abstract:This paper analyses LightGCN in the context of graph recommendation algorithms. Despite the initial design of Graph Convolutional Networks for graph classification, the non-linear operations are not always essential. LightGCN enables linear propagation of embeddings, enhancing performance. We reproduce the original findings, assess LightGCN’s robustness on diverse datasets and metrics, and explore Graph Diffusion as an augmentation of signal propagation in LightGCN.