ListarGrupo de Visión Artificial e Recoñecemento de Patróns (VARPA) por tema "Gradient descent"
Mostrando ítems 1-1 de 1
-
Multi-Adaptive Optimization for multi-task learning with deep neural networks
(Elsevier B.V., 2024-02)[Abstract]: Multi-task learning is a promising paradigm to leverage task interrelations during the training of deep neural networks. A key challenge in the training of multi-task networks is to adequately balance the ...