Carbon Intensity

Carbon intensity, the amount of carbon dioxide emitted per unit of energy or activity, is a key metric for assessing and mitigating the environmental impact of various sectors, particularly computing. Current research focuses on reducing carbon intensity in data centers and AI model training through techniques like optimizing workload scheduling across geographically diverse locations with varying carbon intensities, employing reinforcement learning to manage energy consumption and renewable energy integration, and developing carbon-aware algorithms for federated learning. These efforts leverage predictive models, including neural ordinary differential equations and conformal prediction, to account for temporal and spatial uncertainties in carbon intensity forecasting, ultimately aiming to minimize the environmental footprint of digital technologies.