Quantum Matter Seminar

The standard quantum limit on coherence of laser light was first obtained by Schawlow and Townes in 1958. Except for a small modification in 1999, which decreased this limit by a factor of two, the Schawlow-Townes limit has stood as the ultimate theoretical bound on laser linewidth for 62 years. I will present our theoretical blueprint for building a microwave laser (a maser) with coherence that is better than the standard quantum limit by a factor equal to the number of photons in the laser cavity. The key to our design is a pair of non-linear couplers made of an inductively shunted Josephson junctions that regulate the flow of photons from the gain media (made of a pair of superconducting qubits) to the resonator and out into the transmission line. (This work is related to a concurrent work by Baker, Saadatmand, Berry, and Wiseman).

I will also describe my latest work on machine learning electronic structure of molecules. Over the past few years there has been significant progress toward figuring out how to use machine learning to predict electronic structure, thus avoiding expensive electronic structure calculations. However, all these efforts have focused on single-electron properties. I will argue that 1- and 2-electron reduced density matrices (RDMs) are sensible objects for encoding electronic structure for machine learning. The 2-RDM is especially interesting as it contains sufficient information on electron-electron correlations to compute many observables, including energy, with no additional approximation. I will demonstrate the feasibility of learning 1- and 2-RDMs on the toy problem of linear polymers.