Embedded Files
Graph Neural Networks:
Graph Neural Networks:
Modern Models and Perspectives
Modern Models and Perspectives
In a large number of application domains, the information is naturally represented by graphs and consists of a network of patterns connected by relationships. Graph Neural Networks (GNNs) are a class of machine learning models that can process such information as a whole considering at the same time patterns and relationships. From the first models, GNNs have rapidly evolved exploiting ideas available in other deep learning architectures and ad hoc solutions. However, the peculiarity of graph processing gives rise to new challenging problems, which have been only partially faced. In this talk, modern GNNs will be introduced pointing out the main ideas involved in the evolution of those models and the current applications. Moreover, current research trends will be discussed with theoretical and application challenges. Particular attention will be dedicated to the perspectives due to the peculiarities of the research field.
Franco Scarselli
Full Professor @ Unisi
Google scholar profile: https://scholar.google.it/citations?user=MdCY3T8AAAAJ&hl=it.
Franco Scarselli received the Laurea degree in Computer Science from the University of Pisa, and the PhD degree in Computer Science and Automation Engineering from the University of Florence. In 1999, he moved to the University of Siena, where he is currently a full professor.
Franco Scarselli has been involved as partner and principal investigator in more than 20 research projects focused on theory and applications of machine learning, founded by the Italian Ministry of Education, the Australian Research Council, the Macau Ministry, and private companies. Franco Scarselli is currently associate editor of IEEE TNNLS. He has been co-organizer of several events, including workshops/conferences at IJCNN, GNNNet, ESANN, KES, and editor for special issues and books on Neurocomputing and Springer books.
The research of Franco Scarselli is in the field of machine learning, with a focus on graph neural networks, deep learning, and generative models. His applicative interests include computer vision, and bioinformatics. His main contributions include the definition and the study of the first GNN model, a theoretical analysis of the role of depth in neural networks, a study of the properties of the pagerank algorithms, and several applications of neural networks to image analysis and bioinformatics.
Page updated
Report abuse