Integrated Approach to Controlling Noise on Quantum Devices Towards Optimal Qubit Embedding

Integrated Control Of Noise On Quantum Devices For Optimal Qubit Embedding Quantum Computing holds great promise to solve problems that are currently computationally intractable, and to accelerate the solution of other computationally challenging tasks. However, despite the tremendous progress that has been made over the past few years, quantum computing is still in its infancy. The maturity of quantum computing today is analogous to the early days of transistorized computers when the TRADIC (TRansistorized Airborne DIgital Computer) was developed for the Air Force by Bell Labs in 1954. While TRADIC was a huge step ahead for flyable computers specifically and for computing in general, computation as we know it today was just getting started. For instance, TRADIC used a Mylar sheet with punched holes for its programming. In the last ~70 years, we have come a long way to today’s sophisticated software stacks that have enabled classical computing to revolutionize the world. The role of a robust software stack including operating systems, compilers, programming environments and debugging tools is critical in enabling the computing revolution that we have experienced. For quantum computing to realize the promise and potential needed for the Air Force and DoD broadly, we must learn from the history of computing and shorten the timeline of making quantum computing broadly accessible for a wide array of developers and not just specialists in the field.  Problem Statement: current key impediments to quantum computing.  1. Noisy quantum devices with high error rates: These error rates are due to multiple factors: qubits can only stay in a mixed state for a certain period of time, the entangling and rotational operations are susceptible to accuracy errors, and subtle changes in the surrounding environment can affect qubit operation accuracy. Reducing error rates of the executions is essential for maximizing reliability of the results produced by quantum computers. There is a lack of programming models, compilers and debugging tools that can account for quantum characteristics of noise and decoherence.  2. Unique noise models per device that impede hardware agnostic software stacks: Current quantum devices have unique noise models that require a software stack that is closely coupled to the specific hardware type. Quantum developers lack “write once, run anywhere” paradigms that enable fast innovation in classical computing. Quantum computing software stack has the added complication of time and environment dependent noise and decoherence.  3. Sub-optimal hybrid classical and quantum paradigms: There are several promising quantum algorithms that have extensive use cases in combinatorics, optimization, machine learning and more. However, these algorithms are somewhat limited by the inability of current approaches to use either classical or quantum devices in the near term to find the right set of parameters to make them viable for realistic use cases.