Research project: Kuprov: Neural network analysis of quantum processes
The Spin Dynamics group at the University of Southampton has recently created a family of neural networks that recovered molecular distance distributions from magnetic spectroscopy data (publication). The networks reliably returned distance data for proteins, nucleic acids, photosynthetic reaction centres, and other systems of current chemical and biological interest. The principal problem is that… this was not supposed to be possible: the mathematical transformations in question are ill-posed and numerically unstable. At the moment, it is unclear how those neural networks accomplish what is technically, without regularisation, an impossible mathematical operation.