TORONTO — A novel algorithm developed by IBM scientists is improving the understanding of complex chemical reactions and optimizing quantum computing.
The scientists have developed a new approach to simulate molecules on a quantum computer using a seven-qubit quantum processor to address the molecular structure problem for beryllium hydride (BeH2), which is the largest molecule simulated on a quantum computer to date, according to IBM. The results are significant as they could lead to practical applications such as the creation of novel materials, development of personalized drugs and discovery of more efficient and sustainable energy sources.
In a telephone interview with EE Times, IBM quantum computer research team member Abhinav Kandala outlined how they implemented an algorithm that is efficient with respect to the number of quantum operations required for the simulation. Using six qubits of a seven-qubit processor, they were able to measure BeH2's lowest energy state, a key measurement for understanding chemical reactions. The results were just published in the peer-reviewed journal Nature, which Kandala co-authored.
The research team's paper outlined how it demonstrated the experimental optimization of up to six-qubit Hamiltonian problems with more than one hundred Pauli terms, determining the ground state energy for molecules of increasing size, up to BeH2. This was enabled by “a hardware-efficient quantum optimizer with trial states specifically tailored to the available interactions in our quantum processor, combined with a compact encoding of fermionic Hamiltonians and a robust stochastic optimization routine."
IBM scientists have developed a new approach to simulate molecules on a quantum computer and successfully used six qubits on a purpose-built seven-qubit quantum processor to address the molecular structure problem for beryllium hydride (BeH2) – the largest molecule simulated on a quantum computer to date.
Although this model of BeH2 can be simulated on a “classical" computer, this new approach has the potential to scale towards investigating larger molecules that would traditionally be seen to be beyond the scope of classical computational methods as more powerful quantum systems get built. Kandala said the experiments ultimately show that a hardware-efficient optimizer implemented on a six-qubit superconducting quantum processor can address molecular problems beyond period 1 elements, up to BeH2.
Essentially, he said, they turned the traditional approach of forcing previously known classical computing methods onto quantum hardware on its head: They reversed it by building an algorithm suited to the capability of the current available quantum devices. This allows for extracting the maximal quantum computational power to solve problems that grow exponentially more difficult for classical computers. To characterize the computational power, IBM has adopted a new metric, Quantum Volume. It accounts for the number and quality of qubits, circuit connectivity, and error rates of operations.
Solving problems in chemistry using traditional computing methods presents several problems that could be potentially solved with quantum computing. “All of chemistry today deals with approximate methods," said Kandala. “Trying to solve a problem like this on a classical computer is one that has exponential cost."
The problem scales in complexity with the number of orbitals. A molecular orbital is a mathematical function describing the wave-like behavior of an electron in a molecule. “The hope with quantum computing is to deal with the problem in a precise manner," Kandala said.
If quantum computers are to be used to solve real problems they will need to explore a large space of quantum states. The error rate is just as important as the number of qubits.
For example, a simulation of the simplest molecule, hydrogen, maps four orbitals onto two qubits. “As you try to address larger molecules you have more orbitals you need to account for because you have more orbitals, you need more qubits," said Kandala. “These problems are sufficiently small that you can solve. The reason we are able to attempt to solve problems like this on a quantum computer is because there are mathematical mappings," he said. “The number of orbitals in molecules relates to the number of qubits you require in the simulation."
Part of the point of the experiment is it's an opportunity to compare the quantum computing results with a traditional computing approach, said Kandala, and identify errors. “The hope is to get more information beyond the scope of classical computing," he said. "The field is pretty new."
To help showcase how quantum computers are adept to simulating molecules, developers and users of the IBM Q experience are now able to access a quantum chemistry Jupyter Notebook, which are open source available through the open access QISKit github repo. The IBM Q experience was launched a year ago by placing a robust five-qubit quantum computer on the cloud for anyone to freely access, and most recently upgraded to a 16-qubit processor available for beta access.
“We want to build a community," said Kandala. “We want to learn ourselves, but we want other people to learn too."
—Gary Hilson is a general contributing editor with a focus on memory and flash technologies for EE Times.