ENN: a neural network with DCT adaptive activation functions

The expressiveness of neural networks highly depends on the nature of the activation function, although these are usually assumed predefined and fixed during the training stage. Under a signal processing perspective, in this paper we present Expressive Neural Network (ENN), a novel model in which th...

Descripción completa

Detalles Bibliográficos
Autores: Martínez Gost, Marc|||0000-0003-0070-6807, Pérez Neira, Ana Isabel|||0000-0003-4281-3934, Lagunas Hernandez, Miguel A.|||0000-0003-3338-244X
Tipo de recurso: artículo
Fecha de publicación:2024
País:España
Institución:Universitat Politècnica de Catalunya (UPC)
Repositorio:UPCommons. Portal del coneixement obert de la UPC
Idioma:inglés
OAI Identifier:oai:upcommons.upc.edu:2117/405515
Acceso en línea:https://hdl.handle.net/2117/405515
https://dx.doi.org/10.1109/JSTSP.2024.3361154
Access Level:acceso abierto
Palabra clave:Signal processing
Neural networks (Computer science)
Machine learning
Adaptive activation functions
Discrete cosine transform
Explainable machine learning
Tractament del senyal
Xarxes neuronals (Informàtica)
Aprenentatge automàtic
Àrees temàtiques de la UPC::Enginyeria de la telecomunicació::Processament del senyal
Descripción
Sumario:The expressiveness of neural networks highly depends on the nature of the activation function, although these are usually assumed predefined and fixed during the training stage. Under a signal processing perspective, in this paper we present Expressive Neural Network (ENN), a novel model in which the non-linear activation functions are modeled using the Discrete Cosine Transform (DCT) and adapted using backpropagation during training. This parametrization keeps the number of trainable parameters low, is appropriate for gradient-based schemes, and adapts to different learning tasks. This is the first non-linear model for activation functions that relies on a signal processing perspective, providing high flexibility and expressiveness to the network. We contribute with insights in the explainability of the network at convergence by recovering the concept of bump, this is, the response of each activation function in the output space. Finally, through exhaustive experiments we show that the model can adapt to classification and regression tasks. The performance of ENN outperforms state of the art benchmarks, providing above a 40% gap in accuracy in some scenarios.