Laser-induced electron diffraction (LIED) is a method coined “molecular selfie” that allows pinpointing of the individual atoms inside a single molecule and captures the dynamics where the atoms move during a reaction with the picometer and attosecond spatiotemporal resolutions. A general difficulty for diffraction-based imaging methods is the necessity to extract the molecular configurations from the measured diffraction patterns, which relies on identifying a global extremum in a multi-dimensional solution space. Here, we show how laser-induced electron diffraction (LIED) techniques combined with convolutional neural networks (CNNs) enable atomic-resolution imaging of the complex chiral molecule Fenchone.
 Liu, X., Amini, K., Sanchez, A. et al. Machine learning for laser-induced electron diffraction imaging of molecular structures. Commun Chem 4, 154 (2021).