Graf Makine Öğrenmesi

Graph Data

1. Graphs as DataNodes, edges, features, and adjacency matrices; node-, edge-, and graph-level prediction tasks.1–2 sa
2. Node EmbeddingsLearning vectors for nodes from random-walk co-occurrence, the graph analogue of word embeddings.1–2 sa
3. Knowledge Graph EmbeddingsEmbedding entities and relations so that true triples score highly; link prediction over knowledge graphs.1–2 sa

Gnns

1. Message Passing and Graph Convolutional NetworksEach layer aggregates neighbour features: the core computation of graph neural networks.1–2 sa
2. Graph Attention and Graph TransformersLearning how much each neighbour matters, and letting attention reach beyond immediate neighbours.1–2 sa
3. Graph Learning Tasks and ApplicationsNode classification, link prediction, and graph classification in molecules, social networks, and knowledge bases.1–2 sa
4. Expressivity and Oversmoothing in GNNsWhat message passing cannot distinguish (the Weisfeiler-Lehman limit) and why deep GNNs blur node features.1–2 sa