Eco-Aware Machine Learning

This chapter explores eco-friendly algorithms in recommendation systems, focusing on the use of graph neural networks and their application in modeling user-item relationships. It introduces a new dataset that includes environmental impact metrics alongside traditional performance indicators and discusses the need for sustainable practices in machine learning. The importance of continual research and the evolution of benchmarking practices centered on environmental sustainability are emphasized throughout the chapter.