The field of graph neural networks (GNNs) is moving towards enhancing logical expressiveness and personalized learning. Recent developments have focused on improving the ability of GNNs to model complex relationships and capture user preferences. This has led to the development of new architectures and techniques, such as manifold-aware aggregation and heterogeneous graph learning, which have shown promising results in various applications. Notably, these advancements have the potential to improve the performance of GNNs in knowledge graph reasoning, argumentation, and text-attributed graph learning. Some noteworthy papers in this area include: GMTRouter, which proposes a personalized LLM router that captures user preferences from few-shot data. GeoGNN, which introduces a manifold-aware mechanism to mitigate semantic drift in text-attributed graphs.