Quantum Computational Science

Mission Statement:

The mission of the Quantum Computational Science Group is to develop ways of using quantum computers to solve complex scientific problems that are beyond the reach of traditional computers.  We develop innovative quantum algorithms, devise efficient and robust implementations, formulate and run benchmarks, and create software for quantum-enhanced computational workflows.

Summary:

The Quantum Computational Science Group performs cutting-edge research spanning many facets of quantum computing, from fundamental theory to applications and algorithms, programming tools and compilers, hardware characterization, error mitigation and fault tolerance, modeling and simulation, and resource analysis. Our unifying goal is to make quantum computing more practical and applicable, both with today’s limited quantum processors and with future large-scale quantum computing systems.  We primarily target problems in material science, chemistry, combinatorial optimization, and machine learning.  Potential applications include the development of advanced materials and chemical processes, optimization of energy production and distribution, and faster, more efficient artificial intelligence.

In recent years much of our work has focused on developing algorithms and methods for today’s small, noisy quantum processors, such as those available through the ORNL Quantum Computing User Program.  Although such devices are not capable of tackling real-world problems, developing proof-of-principle computations for these devices advances the understanding of quantum computing and demonstrates its potential for many different application areas.  We have developed innovative and efficient algorithms for simulating quantum materials, for calculating accurate ground state energies of molecules, for performing combinatorial optimization, and for performing machine learning tasks such as classification, kernel evaluation, and generative modeling.  We have also developed new methods for characterizing correlated errors and for quantifying the power of quantum processors to generate many-qubit entanglement.

We also develop software frameworks and tools for incorporating quantum computing into conventional computational workflows. Building on ORNL’s previously developed XACC programming framework, we are developing intermediate representations and execution engines to enable single-source hybrid quantum-classical computing with a variety of quantum processors and simulators. We lead the development of a software infrastructure to manage quantum computing resources jointly with classical compute nodes in a user facility context.  We also contribute to the development of efficient quantum circuit simulators using tensor networks and other advanced representations.

As we look to the future, we have begun efforts in estimating and reducing the quantum resources needed to reach utility in various application areas. One line of work has examined tradeoffs between different fault-tolerant logical architectures and has revealed opportunities to reduce resources by 10-100x in some cases.

Quantum computing is at an exciting point in its development. Over the next 5-8 years it is hoped to transition from a proof-of-principle technology to a practical tool that brings unprecedented capabilities to world-class computing facilities such those at ORNL. Many challenges in algorithms and software will need to be overcome in order to reach this tantalizing future, and the Quantum Computational Science Group is playing a key role in this effort.