Distributed Machine Learning with Sparse Heterogeneous Data

This increase in data sources has led to applications that are increasingly high-dimensional. To be both statistically and computationally efficient in this setting, it is then important to develop approaches that can exploit the structure within the data.