Quantum computing is expected to solve computational questions that cannot be addressed by existing classical computing methods. It is now accepted that the very first discipline that will be greatly advanced by quantum computers is quantum chemistry.

In 1982, the Nobel Prize-winning physicist Richard Feynman observed that simulating and then analyzing molecules was so difficult for a digital computer as to make it impossible for any practical use. The problem was not that the equations governing such simulations were difficult.

In fact, they were comparatively straightforward, and had already been known for decades. The problem was that most molecules of interest contained hundreds of electrons, and each of these electrons interacted with every other electron in a quantum mechanical fashion - resulting in millions of interactions that even powerful computers could not handle.

To overcome the quantum nature of the equations, Feynman proposed quantum computers, which perform calculations based on the laws of quantum physics, as the ultimate answer. Unfortunately, such precise manipulation of individual quantum objects was far from technically possible. The joke for the past 35 years has been that quantum computing is always ten years away.

In the past few years, what was once a distant dream has slowly become a reality. Not only do quantum computers now exist, millions of programs have been executed via the cloud, and useful applications have started to emerge.

The power of a quantum computer can be roughly estimated by the number of qubits, or quantum bits: each qubit can represent a 1 and 0 state simultaneously. There are a number of promising hardware approaches to quantum computing, including superconducting, ion trap, and topological. Each has advantages and disadvantages, but superconducting has taken an early lead in terms of scalability. Google, IBM, and Intel have each used this approach to fabricate quantum processors ranging from 49 to 72 qubits. Qubit quality has also improved.

The breakthrough by scientists at Cambridge Quantum Computing (CQC) and their partners at JSR Corp was the ability to model multi-reference states of molecules. Multi-reference states are often needed to describe the “excited states” arising when molecules interact.

The reason such modeling is significant is that “classical” digital computers find it virtually impossible to tackle multi-reference states; in many cases, classical computing methods fail not only quantitatively but also qualitatively in the description of the electronic structure of the molecules.

An outstanding problem - and the one recently solved - is to find ways that a quantum computer can run calculations efficiently and with the required chemical accuracy to make a difference in the real world. The program was run on IBM’s 20 qubit processor, as both CQC and JSR are members of the IBM Q Network.

Why is chemistry of such interest? Chemistry is one of the first commercially lucrative applications for a variety of reasons. Researchers hope to discover more energy-efficient materials to be used in batteries or solar panels. There are also environmental benefits: about two percent of the world’s energy supply goes toward fertilizer production, which is known to be grossly inefficient and could be improved by sophisticated chemical analysis.

Finally, there are applications in personalized medicine, with the possibility of predicting how pharmaceuticals would affect individuals based on their genetic makeup. The long-term vision is the ability to design a drug for a particular individual to maximize treatment and minimize side effects.

There were two strategies employed by CQC and JSR Corp that allowed the researchers to make this advance. First, they used CQC’s proprietary compiler to most efficiently convert the computer program into instructions for qubit manipulation. Such efficiency is particularly essential on today’s low-qubit machines, in which every qubit is needed and speed of execution is critical.

Second, they utilized quantum machine learning, a special sub-field of machine learning that uses vector-like amplitudes rather than mere probabilities. The method of quantum machine learning being used is specially designed for low-qubit quantum computers, offloading some of the calculations to conventional processors.

The next few years will see a dramatic advance in both quantum hardware and software. As calculations become more refined, more industries will be able to take advantage of applications including quantum chemistry. The Gartner Report states that within 4 years, 20 percent of corporations will have a budget for quantum computing. Within ten years, it should be an integral component of technology.

Mark Jackson is Scientific Lead of Business Development at Cambridge Quantum Computing. Dr. Jackson draws on this variety of deep expertise in scientific innovations, computational technology and space exploration to engage with the Singularity University community.