The idea of quantum virtual machine is to build software stack based on the abstraction level: In application layer, we perform quantum simulation of topological dynamics and field theories. In quantum-classical hybrid architecture layer, we study the synergy between quantum circuit, tensor-network, and machine learning. At middleware layer, we explore device independent optimization and characterization of QPU and bridge it to firmware layer. Finally, we integrate the software stack with QPU.

Build randomized measurement toolbox with quantum error mitigation to measure entanglement, twist order parameter, topological phase. Apply it to study entanglement dynamics and topological transition. Develop tensor-network based quantum circuit compression method for time evolution. Design quantum-based molecular generator to enable efficient exploration of a nontrivial portion of chemical space.

Universal Quantum Computer Diagnostic & Recommendation System, which based on randomized measurement toolbox, has obtained patent in Taiwan. By quantum circuit compression, one maximizes the quantum advantage of NISQ devices, enabling more complex simulations. Exploring huge configuration space via quantum generative model has become an essential part in quantum-classical hybrid machine learning.