دانلود کتاب استدلال گراف دانش – دیدگاه عصبی- نمادین

Knowledge graphs have received wide attention recently, and have been successfully applied to many domains and downstream tasks. In this book, we aim at providing a novel neuro-symbolic perspective in summarizing recent advances in a variety of knowledge graph reasoning tasks. By bridging the gap between neural networks and symbolic reasoning, this book offers a unique and comprehensive perspective on knowledge graph reasoning. It aims to equip researchers, practitioners, and enthusiasts with the knowledge and tools necessary to tackle the complex challenges posed by knowledge graphs and contribute to the continued advancement of this exciting field. On one hand, knowledge graphs (KGs) are a collection of triples storing facts about entities1 via relations between them, in the form of (head entity, relation, tail entity). For instance, (Mina Miller, LiveIn, USA) is a triple, where “Mina Miller” and “USA” are entities in the KG connected by the relation “LiveIn”. On the other hand, as the name suggests, knowledge graphs are essentially graphs, where nodes denote entities and links denote relationships between those entities. A toy example of KG can be found in Fig. 1.1a, which contains 5 entities, 3 relations, and 4 observed facts denoted as triples (e.g., (Thomas Alva Edison, IsMarriedTo, Mary Stilwell)).