Information Gain and Loss in Quantum Feedback Protocols*

Recent advances in fabrication and control of small systems where quantum fluctuations are dominant over thermal fluctuations allow for novel studies of thermodynamics in the quantum regime. As seen in stochastic thermodynamics, one may extract work from a system by using the information acquired from measurements. In the quantum world, measurement inevitably induces back-action on the state of the system and the amount of information gained evolves stochastically. In particular, information can be ‘lost’ due to measurement back-action, which, owing to superposition and entanglement, has no classical counterpart. Here, we use a superconducting qubit to study “information dynamics” of a quantum system subject to continuous weak measurement. With different state preparations and varying drive protocols, we experimentally investigate the transition between regimes of information loss and gain.