Green AI

Green AI focuses on minimizing the environmental impact of artificial intelligence systems, primarily by reducing energy consumption during model training and inference. Current research emphasizes efficient model architectures (like lightweight CNNs and sparse autoencoders), optimized training strategies (including adaptive backpropagation and data-centric approaches), and the exploration of energy-efficient runtime infrastructures. This field is crucial for promoting sustainable AI development and deployment, impacting both the scientific community's research practices and the broader societal adoption of AI technologies by reducing their carbon footprint.