A new framework uses machine learning to simultaneously predict molecular properties and generate new molecules using only a small amount of data for training ...

In principle, however, these properties can be seen everywhere. In molecules, the wave properties of atomic nuclei are obvious and can be observed in the internal vibrations of the atomic nuclei ...

The illustration depicts some of the innovative ideas that underpin the new machine learning model that can quickly and accurately predict the dielectric function of simple molecules, such as shifting ...