Playing the Lottery With Concave Regularizers for Sparse Trainable Neural Networks.

The design of sparse neural networks, i.e., of networks with a reduced number of parameters, has been attracting increasing research attention in the last few years.

Age structure in SIRD models for the COVID-19 pandemic-A case study on Italy data and effects on mortality.

The COVID-19 pandemic is bringing disruptive effects on the healthcare systems, economy and social life of countries all over the world. Even though the elder portion of the population is the most severely affected by the COVID-19 disease, the counter-measures introduced so far by governments took into little account the age structure, with restrictions that act uniformly on the population irrespectively of age. In this paper, we introduce a SIRD model with age classes for studying the impact on the epidemic evolution of lockdown policies applied heterogeneously on the different age groups of the population.

Multiclass Sparse Centroids With Application to Fast Time Series Classification.

In this article, we propose an efficient multiclass classification scheme based on sparse centroids classifiers. The proposed strategy exhibits linear complexity with respect to both the number of classes and the cardinality of the feature space. The classifier we introduce is based on binary space partitioning, performed by a decision tree where the assignation law at each node is defined via a sparse centroid classifier.

RAN-GNNs: Breaking the Capacity Limits of Graph Neural Networks.

Graph neural networks (GNNs) have become a staple in problems addressing learning and analysis of data defined over graphs. However, several results suggest an inherent difficulty in extracting better performance by increasing the number of layers. Recent works attribute this to a phenomenon peculiar to the extraction of node features in graph-based tasks, i.

Sparse ℓ- and ℓ-Center Classifiers.

In this article, we discuss two novel sparse versions of the classical nearest-centroid classifier. The proposed sparse classifiers are based on l and l distance criteria, respectively, and perform simultaneous feature selection and classification, by detecting the features that are most relevant for the classification purpose. We formally prove that the training of the proposed sparse models, with both distance criteria, can be performed exactly (i.