Multimodal Graph Benchmark (MM-GRAPH)
收藏资源简介:
Multimodal Graph Benchmark (MM-GRAPH) 是由密歇根大学创建的一个综合性的多模态图数据集,旨在解决现有图学习基准在处理节点丰富语义信息方面的不足。该数据集包含五个不同规模的图学习数据集,适用于多种学习任务,如节点分类和链接预测。每个数据集都包含文本和视觉信息,以更全面地评估图学习算法在真实场景中的表现。MM-GRAPH 的创建过程涉及从真实世界应用中提取数据,并确保数据集的多样性和实用性。该数据集的应用领域广泛,包括社交网络分析、生物系统和推荐系统等,旨在提高依赖多模态图数据的实际应用性能。
Multimodal Graph Benchmark (MM-GRAPH) is a comprehensive multimodal graph dataset developed by the University of Michigan, which aims to address the limitations of existing graph learning benchmarks in processing the rich semantic information of nodes. This dataset encompasses five graph learning datasets of varying scales, supporting multiple learning tasks such as node classification and link prediction. Each dataset incorporates both textual and visual information, enabling a more comprehensive assessment of graph learning algorithms' performance in real-world scenarios. The creation of MM-GRAPH involves extracting data from real-world applications while ensuring the dataset's diversity and practicality. This benchmark has a wide array of application areas, including social network analysis, biological systems, recommendation systems and more, with the goal of improving the performance of real-world applications that rely on multimodal graph data.




