A group of international researchers have used AI to predict the complex chemical interactions of the molecule malonaldehyde. They did this by training an algorithm using a small sample set of the behaviours of the molecule.
This is another example of how machine learning can complete a wide range of tasks when the system is given enough data. Predicting the chemical behaviour of malonaldehyde usually requires several complex quantum mechanical calculations so this breakthrough is a big step into making this process more efficient.
There are potential applications of this in the development of pharmaceuticals and in helping to improve the performance of batteries, solar cells and digital displays.