Латентная Семантическая Структура

AbstractGraph convolutional networks (GCNs) have been applied recently to text classification and produced an excellent performance. However, existing GCN-based methods do not assume an explicit latent semantic structure of documents, …...