4 papers
Compressing deep quaternion neural networks with targeted regularization
Riccardo Vecchi, Simone Scardapane, Danilo Comminiello +1
In recent years, hyper-complex deep networks (such as complex-valued and quaternion-valued neural networks) have received a renewed interest in the literature. They find applicatio…
Widely Linear Kernels for Complex-Valued Kernel Activation Functions
Simone Scardapane, Steven Van Vaerenbergh, Danilo Comminiello +1
Complex-valued neural networks (CVNNs) have been shown to be powerful nonlinear approximators when the input data can be properly modeled in the complex domain. One of the major ch…
Recurrent Neural Networks with Flexible Gates using Kernel Activation Functions
Simone Scardapane, Steven Van Vaerenbergh, Danilo Comminiello +2
Gated recurrent neural networks have achieved remarkable results in the analysis of sequential data. Inside these networks, gates are used to control the flow of information, allow…
Improving Graph Convolutional Networks with Non-Parametric Activation Functions
Simone Scardapane, Steven Van Vaerenbergh, Danilo Comminiello +1
Graph neural networks (GNNs) are a class of neural networks that allow to efficiently perform inference on data that is associated to a graph structure, such as, e.g., citation net…