Kavli Affiliate: Debanjan Chowdhury
| Summary:
Cooling to ground states of many-body Hamiltonians by imaginary-time evolution is central to quantum simulation, but its non-unitary aspect prevents direct implementation on a quantum computer. Existing routes restore it by importing structure from outside the intrinsic quantum dynamics, e.g. by parameter estimation on many copies or recursive compilation. Here, we show that a single randomized snapshot of the quantum many-body system is sufficient to steer the state along the cooling direction on average. Our algorithm, shadow chain, uses randomized measurements to collapse the state to its classical record, which fixes the next unitary update. Snapshots across steps form a trajectory that is a quantum-executable sample of an auxiliary-field path integral; fluctuations among trajectories are exponentially suppressed in system size. With the randomness supplied intrinsically by the Born measurement rule, the trajectories sample themselves rather than going through external accept-reject steps. The resulting algorithm is ansatz-free and exact upon extrapolation to the continuous limit, with the exponential cost confined only to system size rather than temperature.
| Search Query: arXiv Query: search_query=au:”Chowdhury Debanjan”&id_list=&start=0&max_results=10
Read More
RECENT NON-PEER REVIEWED REPORTS FROM KAVLI INSTITUTE FACULTY AND AFFILIATES