[NOVA Math] Statistics and Risk Management Seminar

Abstract1: As the Internet experiences exponential growth and usage, the power consumption of data centers has increased considerably. The electricity consumption in global data centers range between 1.1% and 1.5% of total electricity usage. In this way, renewable energy sources present a promising solution to mitigate the environmental footprint of data center operations. The focus of this study consists of combining a prediciton and an optimization framework regarding the energy consumption of green data centers. We will present the proposed plan of this study, as well as some state of the art approaches that can be useful in a near future.

Abstract2: This study focuses on modelling complex spatio-temporal fire data using marked spatiotemporal point processes, extreme value theory and Deep Learning methods under a Bayesian framework. By working with both, forest and urban fires in Portugal, the goal is to model and compare the dynamics of events between the two systems. Another goal concerns the definition of extreme/non-extreme events for forest and urban areas and employing the Integrated Nested Laplace Approximation (INLA) and Deep Learning algorithms to maximize the accuracy.

In conclusion, the study aims to provide a comprehensive framework for modeling and analyzing complex spatio-temporal fire data, to promote better fire management practices and improve overall public safety.