The field of recommender systems is moving towards addressing issues of fairness and data sparsity. Researchers are exploring innovative methods to counteract popularity bias and enhance recommendation accuracy. One notable direction is the development of regularization-based loss functions and knowledge-free augmentation frameworks that can improve fairness and semantic diversity without relying on external knowledge. Another area of focus is the use of large language models and cross-modal adversarial fusion to enrich interaction data and mitigate distributional shift. These advancements have the potential to significantly improve the performance and reliability of recommender systems. Noteworthy papers include: PBiLoss, which proposes a novel regularization-based loss function to counteract popularity bias, and NodeDiffRec, which enables fine-grained node-level graph generation for recommendations. Additionally, AUV-Fusion introduces a cross-modal adversarial attack framework that adopts high-order user preference modeling and cross-modal adversary generation to enhance the exposure of target items.