A detector that can learn light's fingerprint

The polarization, wavelength and power of a light wave can be simultaneously identified by a compact device made from twisted layers of carbon atoms - with a little help from an artificial neural network. The problem is exacerbated in moiré materials, whose electronic structure can be affected by minute changes in strain and twist angle. [...]although the device is sensitive to multiple light characteristics, practical application of this capability is hampered by the complicated variations in the photovoltage. [...]it would be worthwhile to explore similar schemes with other moiré materials in which large nonlinear responses to applied voltage have recently been uncovered9.