Title: Quantum machine learning using atom-in-molecule-based fragments selected on-the-fly

Abstract: First principles based exploration of chemical space deepens our understanding of chemistry, and might help with the design of new materials or experiments. Due to the computational cost of quantum chemistry methods and the immens number of theoretically possible stable compounds comprehensive in-silico screening remains prohibitive. To overcome this challenge, we combine atoms-in-molecules based fragments, dubbed "amons" (A), with active learning in transferable quantum machine learning (ML) models. The efficiency, accuracy, scalability, and transferability of resulting AML models is demonstrated for important molecular quantum properties, such as energies, forces, atomic charges NMR shifts, polarizabilities, and for systems ranging from organic molecules over 2D materials and water clusters to Watson-Crick DNA base-pairs and even ubiquitin. Conceptually, the AML approach extends Mendeleev's table to effectively account for chemical environments, which allows the systematic reconstruction of many chemistries from local building blocks.