Quantum computers are getting bigger and bigger, but practical ways to harness the computing power they stand out are still few and far between. Researchers have tried simple simulations on existing quantum computers, such as Google's quantum processor Sycamore, which has done simulations of chemical bonds, high-temperature superconductivity and nanowires, and even simulated phases of exotic matter such as time crystals.
On the other hand, scientists are also becoming more aware of the limitations of classical computers, especially when the laws of quantum mechanics are translated into programs that can be run by classical computers, and there is often a "complete out-of-proportion" of the running time or amount of memory from the size of the simulated physical system.
To overcome these obstacles, the researchers began designing special algorithms that made the transition from classical to quantum computers simpler and more natural. Recently, in a new study published in Nature, a team of researchers unveiled a new algorithm that can effectively reduce the statistical errors generated by qubits when processing chemical equations and reduce noise.
The algorithm, developed by chemists and researchers at quantum computers, used up to 16 qubits on Google's 53-qubit quantum computer plane to calculate ground-state energy, the lowest energy state of a molecule. This is the largest quantum chemical calculation ever performed on a true quantum device.
Classical-quantum hybrid computing
Ground-state energy is influenced by various variables, such as the number of electrons in a molecule, the direction in which they spin, and the path they take around the nucleus. This electron energy is encoded in the Schrödinger equation. As the molecules grew larger, the difficulty of solving this equation on classical computers grew exponentially.
In principle, quantum computers have the ability to handle calculations that vary exponentially in complexity and mass, such as those required to solve Schrödinger's equations, because their qubits take advantage of the unique advantages of quantum states. Unlike bits composed of 1s or 0s, qubits can exist in the superposition state of both 0 and 1. But the problem is that qubits are also very fragile and prone to error. The more qubits are used, the less accurate the final answer becomes.
The new algorithm in the new study leverages the combined power of classical computers and quantum computers to solve chemical equations more efficiently and minimize quantum computer errors.
This algorithm uses quantum Monte Carlo techniques, which are a method system for calculating probabilities when there are a large number of unknown random variables at work. The researchers used their algorithm to determine the ground-state energy of three molecules, including calculations of H using 8 qubits, calculations of N molecules using 12 qubits, and 16 qubits of energy for two carbon atoms in a solid diamond crystal.
A classical computer could handle most of the quantum Monte Carlo simulations in the algorithm, and Sycamore joined the most complex step in the calculation: calculating the overlap between the experimental wave function (a guess of the mathematical description of the ground state energy performed by a quantum computer) and the sample wave function, which is part of the Monte Carlo statistical process. This overlap provides a set of constraints for Monte Carlo sampling, the so-called boundary conditions, which ensure the statistical efficiency of the calculation.
This division of labor between classical computers and quantum processors helped the team make the best use of both resources. Studies have shown that even using a low-resolution approximation of the ground state (just a few qubits encoding the position of an electron) on a quantum computer, a classical computer can efficiently solve a higher resolution version (the result of a more realistic electron position).
Perfect the chemistry of quantum computing
Previous records for solving ground-state energies have used 12 qubits and a method called the Variational Quantum Solver (VQE). But VQE ignores the effects of interacting electrons, which is actually an important variable in calculating ground-state energy, and the quantum Monte Carlo algorithm in the new study includes this variable.
Classical-quantum hybrid computing in this new work is as precise as some of the best classical methods. This suggests that the use of quantum computers can help solve problems more accurately and quickly, a key milestone in quantum computing.
The researchers believe that adding virtual correlation techniques to classical computers could help chemists process larger molecules. They hope to continue to tweak algorithms to make computing more efficient, while engineers will also work to build better quantum hardware, using the tools they already have and the most advanced quantum information science to perfect the chemistry of quantum computing.
Mastering the ability to accurately calculate ground-state energy can help chemists develop new materials. The application prospects of this algorithm are very broad, such as designing materials, accelerating nitrogen fixation in agriculture, hydrolysis for clean energy, etc., to achieve more sustainable goals.
#创作团队:
Compile: Takeko
Typography: Wenwen
#参考来源:
https://news.columbia.edu/news/toward-quantum-computer-calculates-molecular-energy
https://ai.googleblog.com/2022/03/hybrid-quantum-algorithms-for-quantum.html
#图片来源:
Cover image: Nicoletta Barolini/Columbia
首图：Erik Lucero, Research Scientist and Lead Production Quantum Hardware