Machine Learning Models Tackle Global Aid Data and Industrial Waste
Researchers are applying machine learning to unexpected areas, from sorting $2.8 trillion in international development aid data to identifying reusable machine parts to reduce industrial waste.
Researchers at ETH Zurich and LMU Munich apply machine learning to international development aid, training a model on 20 years of project data totaling $2.8 trillion. The algorithm analyzes 3.2 million projects across 200 dimensions, accomplishing a sorting task that is entirely impossible for human beings. Initial findings from this massive dataset reveal that spending on inclusion and diversity increases while climate spending surprisingly decreases in recent years.
Separately, German R&D organization Fraunhofer develops a machine learning system to identify and sort industrial machine parts that would otherwise head to the scrap yard. Because parts often look similar but differ mechanically, or look different due to rust but are functionally identical, the system goes beyond simple camera views. It incorporates weight measurements, 3D camera scans, and metadata to accurately suggest what each component is.
This AI-powered identification process saves human inspectors from starting their analysis from scratch, significantly accelerating the processing of millions of components. The technology aims to save tens of thousands of parts by ensuring they are reused or recycled responsibly instead of being wasted. Both projects demonstrate how artificial intelligence expands into unexpected niches by sorting through enormous volumes of complex data.