Distributed Learning
This line aims to develop models and algorithms that allow data collected at the edge to be aggregated and processed in a distributed manner, using distributed or centralized learning mechanisms, and taking advantage of a hierarchical design of the computing infrastructure that includes computing at the edge to provide sufficient processing and storage capacity, as well as adequate response times. Decision making about offloading, that is, data transfer and processing to remote devices, must be carried out based on the availability of data at different levels of the computational hierarchy, as well as the context of use and the current environment in which the user is located.