During the project, we designed and developed a prototype software library that provides a unified, user-friendly, and well-documented environment for hybrid quantum–classical simulations. The software includes illustrative examples demonstrating its accessibility and practical use for researchers.
A key technical achievement was the establishment of a standardized and abstract representation for expressing quantum computations, such as evolving an initial state in time and measuring correlation functions as an example from condensed matter physics. This representation allows users to define high-level simulation tasks independently of the (classical or quantum) compute resources that will eventually be used.
This is complemented by a toolbox of transformations that can use classical compute, such as the Density Matrix Renormalization Group (DMRG) for ground state search, Matrix Product Operator (MPO)-based time evolution, as well as state-of-the-art variational and gradient-based algorithms for gate-based preparation of tensor network states on a quantum device.
The user can then either
(i) use only classical compute and evaluate the entire simulation, e.g. as a sandbox developing tool,
(ii) use minimal classical compute and translate every step to concrete quantum gates
(iii) use classical compute adaptively, to compress parts of a simulation and achieve a shallower circuit on the quantum device. This can drastically reduce the quantum resources, as well as fidelity requirements.
The prototype successfully demonstrated these capabilities through a suite of representative examples, validating the framework’s potential to streamline the design, testing, and execution of quantum simulations across different computational regimes.
The prototype is currently closed source, but most of the code base will eventually be open source, once a business model for the project's future is defined.
