Paris-based Pasqal, a full-stack neutral atom quantum computing provider, and BASF, the multinational chemical juggernaut, are announcing a new partnership focused on weather modeling and other computational fluid dynamics applications. The problem space relies on solving complex nonlinear differential equations, a task for which, it turns out, Pasqal’s hardware and algorithms are especially well-suited.
Neutral atoms and physics-informed neural networks
Benno Broer, Pasqal’s Chief Commercial Officer, who was CEO of Qu & Co. which was focused on quantum algorithms and software, and merged with the erstwhile hardware-focused Pasqal in January of this year.
Broer explained that Pasqal’s hardware platform implements qubits (quantum bits), using individually trapped atoms that are manipulated with laser beams, and that the company produces 100-qubit systems today. The neutral atom platform supports something called “analog mode,” which enables addressing all of those qubits concurrently, thereby enabling an important quantum computing behavior called entanglement, where multiple qubits act as a single system and influence each other.
Building on this, Pasqal’s algorithm technology can implement quantum neural networks, the quantum computing equivalent of physics-informed neural networks (PINNs), a subset of physics-informed machine learning (PIML). In the PIML world, models can be trained using a combination of data and equations that describe the laws of physics underlying the modeled phenomena.
Differential equations, weather and climate
PIML techniques can be used to solve differential equations, which is the key to attacking computational fluid dynamics applications, including weather modeling. According to Pasqal’s press release, BASF can then use “parameters generated by the weather models to simulate crop yields and growth stages, as well as to predict drift when applying crop protection products.”
The weather modeling further serves BASF’s digital farming product portfolio, including an advanced crop optimization platform. This takes quantum computing down from the ivory tower, and applies it, quite literally, in the field.
Perhaps even more intriguing, Broer told VentureBeat the equations used to model short-term weather patterns and those for long-term climate modeling are, in fact, similar. Scaling up the time dimension can allow the technology being applied to weather modeling in the near future to be applicable to climate modeling later, and perhaps be used to mitigate the effects of climate change.
Given the heatwaves impacting so many regions across the world this week, even the potential of quantum computing to help mitigate climate change impact is good news indeed. And if we’re going to “tech our way out of this” (the phrase attributed to Kleiner Perkins chairman John Doerr), then an approach that combines quantum computing and physics-informed machine learning seems like a good start.