Opening Quantum Computing To Chemists With Openfermion

Published by whurley (10/26/2017)

Google’s Release of OpenFermion Will Help Drive Chemistry into the Quantum Age

The incredible work scientists are doing in quantum computing blows me away every day. On Monday, Google announced the release of OpenFermion, a new open-source platform for chemists to compile, analyze and simulate chemistry and material science problems on a quantum computer.  OpenFermion is a Python package that Google designed to be quantum programming language agnostic. It will be compatible with different quantum computer hardware implementations. Harvard, ETH Zurich, NASA, Lawrence Berkeley National Labs, University of Michigan, Dartmouth College, Oxford University, and Rigetti Computing all contributed to the project.  An impressive list, to say the least.

Rigetti Computing summed up the announcement this way:

“OpenFermion helps users transform quantum chemistry simulations into quantum computations with simple methods and familiar data structures. It integrates with common quantum chemistry simulation packages, such as Psi4 and PySCF, along with quantum computation environments such as Rigetti’s Forest. This combination means OpenFermion make it easier for scientists in either field to develop methods for simulating Fermionic systems using quantum devices.”

Two Worlds Collided, and You Were There

Quantum chemistry represents a good test for early quantum computers. There are molecular systems which have relevant chemical applications and relatively manageable sizes. The team addressed this subject in a first-rate paper about the project entitled “OpenFermion: The Electronic Structure Package for Quantum Computers.”

The paper states:

“Unfortunately, developing algorithms in this area can require a prohibitive amount of domain expertise. For example, quantum algorithms experts may find the chemistry literature rife with jargon and unknown approximations while chemists find themselves unfamiliar with the concepts used in quantum information. As has been seen though, both have a crucial role to play in developing algorithms for these emerging devices.”

In other words, you need two PhDs to pull this off. That, or a kick-a$$, Google-scale rolodex.

And You’re Telling Me This Because?

Google envisions OpenFermion as a way to bridge this gap. Google also hopes OpenFermion becomes a “community standard for putting quantum chemistry on quantum computers.”  In terms of it being platform agnostic, open source, and engineered to a high standard, I sure hope so.

I think in the grand scheme of quantum computing, the audience for this package may be relatively small.  In the near term, researchers will be compiling to a limited simulator.  But when you consider that we have a functional software package before a quantum computer even exists? Now I’m excited.  It’s like we’re back in the 1920s, with papers coming fast and furious as physicists worked to understand quantum mechanics. Or back in the 1950s as the evolution of the computer was accelerating.

The difference is that we know an ocean more about software development now.  It’s likely we’ll have a library of software ready when the first quantum computer arrives.