Description
Standard descriptions of open quantum systems rely heavily on weak-coupling, Markovian, and slow-driving approximations. In quantum optics, the quantum regression theorem relates single-time Liouvillian evolution to multi-time correlation functions, governing key observables such as emission spectra and photon statistics. In quantum thermodynamics, weak-coupling master equations similarly constrain the analysis of energy fluctuations to low-order work moments. However, quantum systems routinely exist far beyond these limits—where strong system-environment coupling, structured reservoirs (such as vibrational modes in solid-state emitters), and rapid driving induce memory effects that invalidate standard frameworks.
In this talk, I will present two powerful, non-perturbative approaches designed to overcome these limitations across quantum optics and quantum thermodynamics. First, I will introduce filter theory (the sensor method), which incorporates auxiliary detector degrees of freedom to directly compute multi-time correlation functions and optical spectra, completely bypassing the quantum regression theorem. Second, I will present a unified process-tensor framework that maps non-Markovian dynamics onto generalised multi-segment time contours. By integrating physical time evolution with an auxiliary counting field $\chi$ along a single generalised-time axis, this tensor-network methodology enables the numerically exact evaluation of full work counting statistics across arbitrary coupling strengths and driving protocols.
Finally, I will demonstrate how these complementary techniques reveal rich non-equilibrium physics at the intersection of optics and thermodynamics. From resolving fine-grained dressed-state interference in optical emission spectra to exposing microscopic quantum control features in full work distributions that are completely masked by low-order moments, these approaches establish a versatile toolkit for probing thermodynamics and photonics in modern quantum technologies.