It is expected that quantum computing will allow us to solve computational problems that can not be solved by existing classical methods of computation. At present, it is considered that the very first discipline that will receive the most powerful impetus from quantum achievements is quantum chemistry. In 1982, Nobel Prize winning physicist Richard Feynman noticed that simulation, and then analysis of molecules, is so difficult for a digital computer that it becomes almost useless for these cases.
And the problem was not that the equations governing these simulations were complex, no. In fact, they were relatively simple and straightforward, and have known about them for many years. The problem was that most of the molecules of interest to scientists contained hundreds of electrons, and each of these electrons interacted with each other electron in a quantum mechanical way — which led to millions of interactions that even the most powerful computer could not process.
To overcome the quantum nature of these equations, Feynman proposed creating quantum computers that would perform calculations based on the laws of quantum physics as an answer. Unfortunately, such precise manipulation of individual quantum objects was almost impossible from a technical point of view. The joke that many have had to tire over the past 35 years was that quantum computers “will be in ten years.”
Over the past couple of years, what was once a distant dream has slowly turned into reality. Quantum computers are starting to appear, millions of programs are working through the cloud, and useful applications are taking shape.
The power of a quantum computer can be roughly estimated by the number of qubits, or quantum bits: each qubit can represent both 1 and 0 at the same time. There are several promising hardware approaches to quantum computing, including superconductivity, ion traps, and topological transistors. Each of them has pros and cons, but superconductors rank first in terms of scalability. Google, IBM and Intel use this approach to create quantum processors in the range from 49 to 72 qubits. The quality of qubit is also constantly improving.
Breakthrough in chemistry
The breakthrough of scientists from the Cambridge Center for Quantum Computing and their partners from JSR Corp was the ability to simulate multi-reference states of molecules. Multireference states are often necessary to describe the “excited states” that appear when molecules interact.
The reason that such a simulation is important is that classic digital computers can do almost nothing with multi-reference states; in many cases, classical methods of calculation are not only quantitatively but also qualitatively unable to describe the electronic structure of molecules.
An important problem that was recently solved was to find ways in which a quantum computer could efficiently perform calculations and with the required chemical accuracy for the real world. The program was launched on a 20-qubit IBM processor.
Why did chemistry get into the field of such interest? Chemistry is one of the most profitable commercial applications for a number of reasons. Scientists expect to find more energy-efficient materials that can be used in batteries or solar panels. There are also environmental benefits: about two percent of the world’s energy is used to produce fertilizers, which are terribly inefficient and can be improved by complex chemical analysis.
Finally, there are applications in personalized medicine, with the ability to predict how pharmaceuticals will affect people based on their genetics. In the long term, it is possible to develop a drug for a specific person for the most effective treatment and to minimize side effects.
CQC and JSR Corp had two strategies that allowed scientists to accomplish this breakthrough. First, they used their own CQC compiler to most effectively convert a computer program into instructions for manipulating a qubit. Such efficiency is especially important on modern low-kill machines, in which each qubit is important and necessary, and the speed of execution is crucial.
Second, they used quantum machine learning, a special machine learning subfield that uses vector amplitude, and not just probabilities. The used method of quantum machine learning was specially developed for low-hit quantum computers, with partial unloading using traditional processors.
In the next few years, a significant improvement in quantum hardware and software is expected. As calculations become more and more accurate, more and more industries can take advantage of quantum computer applications, including quantum chemistry. Gartner predicts that within four years, 20% of corporations will have a budget for quantum computing. In ten years they will become an integral component of technology.